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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 949 - (show annotations) (download)
Thu Mar 8 10:28:49 2012 UTC (2 years, 5 months ago) by zherczeg
File size: 228803 byte(s)
Rename madvise to posix_madvise
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 19. Added binary file support to pcregrep, including the -a, --binary-files,
73 -I, and --text options.
74
75 20. The madvise function is renamed for posix_madvise for QNX compatibility
76 reasons. Fixed by Giuseppe D'Angelo.
77
78
79 Version 8.30 04-February-2012
80 -----------------------------
81
82 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
83 name is defined in ctype.h.
84
85 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
86 only in quite long subpatterns.
87
88 3. Removed the function pcre_info(), which has been obsolete and deprecated
89 since it was replaced by pcre_fullinfo() in February 2000.
90
91 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
92 match a (*MARK), and the match failed at the start of the subject, a
93 reference to memory before the start of the subject could occur. This bug
94 was introduced by fix 17 of release 8.21.
95
96 5. A reference to an unset group with zero minimum repetition was giving
97 totally wrong answers (in non-JavaScript-compatibility mode). For example,
98 /(another)?(\1?)test/ matched against "hello world test". This bug was
99 introduced in release 8.13.
100
101 6. Add support for 16-bit character strings (a large amount of work involving
102 many changes and refactorings).
103
104 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
105 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
106 from a file.
107
108 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
109 rounding is not applied in this particular case).
110
111 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
112 if they appear, or are escaped, in patterns.
113
114 10. Get rid of a number of -Wunused-but-set-variable warnings.
115
116 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
117 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
118 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
119 also returns the mark "x". This bug applied to capturing parentheses,
120 non-capturing parentheses, and atomic parentheses. It also applied to some
121 assertions.
122
123 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
124 information out of configure.ac instead of relying on pcre.h.generic, which
125 is not stored in the repository.
126
127 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
128 -lreadline.
129
130 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
131
132 15. Applied Graycode's patch to put the top-level frame on the stack rather
133 than the heap when not using the stack for recursion. This gives a
134 performance improvement in many cases when recursion is not deep.
135
136 16. Experimental code added to "pcretest -C" to output the stack frame size.
137
138
139 Version 8.21 12-Dec-2011
140 ------------------------
141
142 1. Updating the JIT compiler.
143
144 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
145 are added as well.
146
147 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
148 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
149 calling _pcre_jit_exec. Some extra comments are added.
150
151 4. (*MARK) settings inside atomic groups that do not contain any capturing
152 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
153 was introduced by change 18 for 8.20.
154
155 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
156 ECMA-262 standard.
157
158 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
159 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
160 This bug was probably introduced by change 9 of 8.13.
161
162 7. While fixing 6 above, I noticed that a number of other items were being
163 incorrectly rejected as "not fixed length". This arose partly because newer
164 opcodes had not been added to the fixed-length checking code. I have (a)
165 corrected the bug and added tests for these items, and (b) arranged for an
166 error to occur if an unknown opcode is encountered while checking for fixed
167 length instead of just assuming "not fixed length". The items that were
168 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
169 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
170 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
171
172 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
173 being incorrectly compiled and would have given unpredicatble results.
174
175 9. A possessively repeated subpattern with minimum repeat count greater than
176 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
177 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
178 the first (A) could occur when it should not.
179
180 10. Add a cast and remove a redundant test from the code.
181
182 11. JIT should use pcre_malloc/pcre_free for allocation.
183
184 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
185 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
186 is anything other than /usr, -L is still shown).
187
188 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
189
190 14. Perl does not support \N without a following name in a [] class; PCRE now
191 also gives an error.
192
193 15. If a forward reference was repeated with an upper limit of around 2000,
194 it caused the error "internal error: overran compiling workspace". The
195 maximum number of forward references (including repeats) was limited by the
196 internal workspace, and dependent on the LINK_SIZE. The code has been
197 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
198 the default depends on LINK_SIZE. There is a new upper limit (for safety)
199 of around 200,000 forward references. While doing this, I also speeded up
200 the filling in of repeated forward references.
201
202 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
203 incorrectly expecting the subject to contain another "a" after the start.
204
205 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
206 in the match, the SKIP should be ignored. This was not happening; instead
207 the SKIP was being treated as NOMATCH. For patterns such as
208 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
209 tested.
210
211 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
212 now much more compatible with Perl, in particular in cases where the result
213 is a non-match for a non-anchored pattern. For example, if
214 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
215 "m", where previously it did not return a name. A side effect of this
216 change is that for partial matches, the last encountered mark name is
217 returned, as for non matches. A number of tests that were previously not
218 Perl-compatible have been moved into the Perl-compatible test files. The
219 refactoring has had the pleasing side effect of removing one argument from
220 the match() function, thus reducing its stack requirements.
221
222 19. If the /S+ option was used in pcretest to study a pattern using JIT,
223 subsequent uses of /S (without +) incorrectly behaved like /S+.
224
225 21. Retrieve executable code size support for the JIT compiler and fixing
226 some warnings.
227
228 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
229 not work when the shorter character appeared right at the end of the
230 subject string.
231
232 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
233 systems.
234
235 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
236 output it when the /M option is used in pcretest.
237
238 25. The CheckMan script was not being included in the distribution. Also, added
239 an explicit "perl" to run Perl scripts from the PrepareRelease script
240 because this is reportedly needed in Windows.
241
242 26. If study data was being save in a file and studying had not found a set of
243 "starts with" bytes for the pattern, the data written to the file (though
244 never used) was taken from uninitialized memory and so caused valgrind to
245 complain.
246
247 27. Updated RunTest.bat as provided by Sheri Pierce.
248
249 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
250
251 29. Computation of memory usage for the table of capturing group names was
252 giving an unnecessarily large value.
253
254
255 Version 8.20 21-Oct-2011
256 ------------------------
257
258 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
259 a POSIX class. After further experiments with Perl, which convinced me that
260 Perl has bugs and confusions, a closing square bracket is no longer allowed
261 in a POSIX name. This bug also affected patterns with classes that started
262 with full stops.
263
264 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
265 captured substring, but while checking the failing first alternative,
266 substring 1 is temporarily captured. If the output vector supplied to
267 pcre_exec() was not big enough for this capture, the yield of the function
268 was still zero ("insufficient space for captured substrings"). This cannot
269 be totally fixed without adding another stack variable, which seems a lot
270 of expense for a edge case. However, I have improved the situation in cases
271 such as /(a)(b)x|abc/ matched against "abc", where the return code
272 indicates that fewer than the maximum number of slots in the ovector have
273 been set.
274
275 3. Related to (2) above: when there are more back references in a pattern than
276 slots in the output vector, pcre_exec() uses temporary memory during
277 matching, and copies in the captures as far as possible afterwards. It was
278 using the entire output vector, but this conflicts with the specification
279 that only 2/3 is used for passing back captured substrings. Now it uses
280 only the first 2/3, for compatibility. This is, of course, another edge
281 case.
282
283 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
284 main code base, and can be used by building with --enable-jit. When this is
285 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
286 runtime --no-jit option is given.
287
288 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
289 ovector, the return from the function was zero, implying that there were
290 other matches that did not fit. The correct "exactly full" value is now
291 returned.
292
293 6. If a subpattern that was called recursively or as a subroutine contained
294 (*PRUNE) or any other control that caused it to give a non-standard return,
295 invalid errors such as "Error -26 (nested recursion at the same subject
296 position)" or even infinite loops could occur.
297
298 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
299 computing the minimum length on reaching *ACCEPT, and so ended up with the
300 wrong value of 1 rather than 0. Further investigation indicates that
301 computing a minimum subject length in the presence of *ACCEPT is difficult
302 (think back references, subroutine calls), and so I have changed the code
303 so that no minimum is registered for a pattern that contains *ACCEPT.
304
305 8. If (*THEN) was present in the first (true) branch of a conditional group,
306 it was not handled as intended. [But see 16 below.]
307
308 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
309 Sheri Pierce.
310
311 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
312 the first byte in a match must be "a".
313
314 11. Change 17 for 8.13 increased the recursion depth for patterns like
315 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
316 pattern contains any instances of (*THEN). If it does not, the old
317 optimizations are restored. It would be nice to do this on a per-group
318 basis, but at the moment that is not feasible.
319
320 12. In some environments, the output of pcretest -C is CRLF terminated. This
321 broke RunTest's code that checks for the link size. A single white space
322 character after the value is now allowed for.
323
324 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
325 For "fr", it uses the Windows-specific input and output files.
326
327 14. If (*THEN) appeared in a group that was called recursively or as a
328 subroutine, it did not work as intended. [But see next item.]
329
330 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
331 pattern fragments (but not containing any | characters). If A and B are
332 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
333 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
334 D. In other words, Perl considers parentheses that do not contain any |
335 characters to be part of a surrounding alternative, whereas PCRE was
336 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
337 differently. PCRE now behaves in the same way as Perl, except in the case
338 of subroutine/recursion calls such as (?1) which have in any case always
339 been different (but PCRE had them first :-).
340
341 16. Related to 15 above: Perl does not treat the | in a conditional group as
342 creating alternatives. Such a group is treated in the same way as an
343 ordinary group without any | characters when processing (*THEN). PCRE has
344 been changed to match Perl's behaviour.
345
346 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
347 RunGrepTest script failed.
348
349 18. Change 22 for version 13 caused atomic groups to use more stack. This is
350 inevitable for groups that contain captures, but it can lead to a lot of
351 stack use in large patterns. The old behaviour has been restored for atomic
352 groups that do not contain any capturing parentheses.
353
354 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
355 suppress the check for a minimum subject length at run time. (If it was
356 given to pcre_exec() or pcre_dfa_exec() it did work.)
357
358 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
359 fail to work when decoding hex characters in data strings in EBCDIC
360 environments.
361
362 21. It appears that in at least one Mac OS environment, the isxdigit() function
363 is implemented as a macro that evaluates to its argument more than once,
364 contravening the C 90 Standard (I haven't checked a later standard). There
365 was an instance in pcretest which caused it to go wrong when processing
366 \x{...} escapes in subject strings. The has been rewritten to avoid using
367 things like p++ in the argument of isxdigit().
368
369
370 Version 8.13 16-Aug-2011
371 ------------------------
372
373 1. The Unicode data tables have been updated to Unicode 6.0.0.
374
375 2. Two minor typos in pcre_internal.h have been fixed.
376
377 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
378 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
379 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
380
381 4. There were a number of related bugs in the code for matching backrefences
382 caselessly in UTF-8 mode when codes for the characters concerned were
383 different numbers of bytes. For example, U+023A and U+2C65 are an upper
384 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
385 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
386 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
387 2-byte code at the end of the subject (it thought there wasn't enough data
388 left).
389
390 5. Comprehensive information about what went wrong is now returned by
391 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
392 as the output vector has at least 2 elements. The offset of the start of
393 the failing character and a reason code are placed in the vector.
394
395 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
396 now returned is for the first byte of the failing character, instead of the
397 last byte inspected. This is an incompatible change, but I hope it is small
398 enough not to be a problem. It makes the returned offset consistent with
399 pcre_exec() and pcre_dfa_exec().
400
401 7. pcretest now gives a text phrase as well as the error number when
402 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
403 failure, the offset and reason code are output.
404
405 8. When \R was used with a maximizing quantifier it failed to skip backwards
406 over a \r\n pair if the subsequent match failed. Instead, it just skipped
407 back over a single character (\n). This seems wrong (because it treated the
408 two characters as a single entity when going forwards), conflicts with the
409 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
410 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
411 has been changed.
412
413 9. Some internal refactoring has changed the processing so that the handling
414 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
415 time (the PCRE_DOTALL option was changed this way some time ago: version
416 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
417 which was always a bit of a fudge. It also means that there is one less
418 argument for the match() function, which reduces its stack requirements
419 slightly. This change also fixes an incompatibility with Perl: the pattern
420 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
421
422 10. More internal refactoring has drastically reduced the number of recursive
423 calls to match() for possessively repeated groups such as (abc)++ when
424 using pcre_exec().
425
426 11. While implementing 10, a number of bugs in the handling of groups were
427 discovered and fixed:
428
429 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
430 (a|)*(?1) gave a compile-time internal error.
431 ((a|)+)+ did not notice that the outer group could match an empty string.
432 (^a|^)+ was not marked as anchored.
433 (.*a|.*)+ was not marked as matching at start or after a newline.
434
435 12. Yet more internal refactoring has removed another argument from the match()
436 function. Special calls to this function are now indicated by setting a
437 value in a variable in the "match data" data block.
438
439 13. Be more explicit in pcre_study() instead of relying on "default" for
440 opcodes that mean there is no starting character; this means that when new
441 ones are added and accidentally left out of pcre_study(), testing should
442 pick them up.
443
444 14. The -s option of pcretest has been documented for ages as being an old
445 synonym of -m (show memory usage). I have changed it to mean "force study
446 for every regex", that is, assume /S for every regex. This is similar to -i
447 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
448 using it. It makes it easier to run collections of tests with and without
449 study enabled, and thereby test pcre_study() more easily. All the standard
450 tests are now run with and without -s (but some patterns can be marked as
451 "never study" - see 20 below).
452
453 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
454 restoration of the capturing data to the outer values was not happening
455 correctly.
456
457 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
458 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
459 pattern had matched an empty string, and so incorrectly returned a no
460 match.
461
462 17. There was optimizing code for the last branch of non-capturing parentheses,
463 and also for the obeyed branch of a conditional subexpression, which used
464 tail recursion to cut down on stack usage. Unfortunately, now that there is
465 the possibility of (*THEN) occurring in these branches, tail recursion is
466 no longer possible because the return has to be checked for (*THEN). These
467 two optimizations have therefore been removed. [But see 8.20/11 above.]
468
469 18. If a pattern containing \R was studied, it was assumed that \R always
470 matched two bytes, thus causing the minimum subject length to be
471 incorrectly computed because \R can also match just one byte.
472
473 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
474 was incorrectly computed.
475
476 20. If /S is present twice on a test pattern in pcretest input, it now
477 *disables* studying, thereby overriding the use of -s on the command line
478 (see 14 above). This is necessary for one or two tests to keep the output
479 identical in both cases.
480
481 21. When (*ACCEPT) was used in an assertion that matched an empty string and
482 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
483
484 22. When an atomic group that contained a capturing parenthesis was
485 successfully matched, but the branch in which it appeared failed, the
486 capturing was not being forgotten if a higher numbered group was later
487 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
488 group 1 to "a", when in fact it should be unset. This applied to multi-
489 branched capturing and non-capturing groups, repeated or not, and also to
490 positive assertions (capturing in negative assertions does not happen
491 in PCRE) and also to nested atomic groups.
492
493 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
494 subject after a captured substring, to make it easier to tell which of a
495 number of identical substrings has been captured.
496
497 24. The way atomic groups are processed by pcre_exec() has been changed so that
498 if they are repeated, backtracking one repetition now resets captured
499 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
500 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
501 "aaa". Previously, it would have been "a". As part of this code
502 refactoring, the way recursive calls are handled has also been changed.
503
504 25. If an assertion condition captured any substrings, they were not passed
505 back unless some other capturing happened later. For example, if
506 (?(?=(a))a) was matched against "a", no capturing was returned.
507
508 26. When studying a pattern that contained subroutine calls or assertions,
509 the code for finding the minimum length of a possible match was handling
510 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
511 group 1 called group 2 while simultaneously a separate group 2 called group
512 1). A stack overflow occurred in this case. I have fixed this by limiting
513 the recursion depth to 10.
514
515 27. Updated RunTest.bat in the distribution to the version supplied by Tom
516 Fortmann. This supports explicit test numbers on the command line, and has
517 argument validation and error reporting.
518
519 28. An instance of \X with an unlimited repeat could fail if at any point the
520 first character it looked at was a mark character.
521
522 29. Some minor code refactoring concerning Unicode properties and scripts
523 should reduce the stack requirement of match() slightly.
524
525 30. Added the '=' option to pcretest to check the setting of unused capturing
526 slots at the end of the pattern, which are documented as being -1, but are
527 not included in the return count.
528
529 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
530 compiled something random. Now it gives a compile-time error (as does
531 Perl).
532
533 32. A *MARK encountered during the processing of a positive assertion is now
534 recorded and passed back (compatible with Perl).
535
536 33. If --only-matching or --colour was set on a pcregrep call whose pattern
537 had alternative anchored branches, the search for a second match in a line
538 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
539 matched the line "0102" twice. The same bug affected patterns that started
540 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
541 twice.
542
543 34. Previously, PCRE did not allow quantification of assertions. However, Perl
544 does, and because of capturing effects, quantifying parenthesized
545 assertions may at times be useful. Quantifiers are now allowed for
546 parenthesized assertions.
547
548 35. A minor code tidy in pcre_compile() when checking options for \R usage.
549
550 36. \g was being checked for fancy things in a character class, when it should
551 just be a literal "g".
552
553 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
554 appearance of a nested POSIX class supersedes an apparent external class.
555 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
556 unescaped square brackets may also appear as part of class names. For
557 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
558 more like Perl. (But see 8.20/1 above.)
559
560 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
561 was because it thought it was \N{name}, which is not supported).
562
563 39. Add minix to OS list not supporting the -S option in pcretest.
564
565 40. PCRE tries to detect cases of infinite recursion at compile time, but it
566 cannot analyze patterns in sufficient detail to catch mutual recursions
567 such as ((?1))((?2)). There is now a runtime test that gives an error if a
568 subgroup is called recursively as a subpattern for a second time at the
569 same position in the subject string. In previous releases this might have
570 been caught by the recursion limit, or it might have run out of stack.
571
572 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
573 happen only once. PCRE was, however incorrectly giving a compile time error
574 "recursive call could loop indefinitely" because it cannot analyze the
575 pattern in sufficient detail. The compile time test no longer happens when
576 PCRE is compiling a conditional subpattern, but actual runaway loops are
577 now caught at runtime (see 40 above).
578
579 42. It seems that Perl allows any characters other than a closing parenthesis
580 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
581 has been changed to be the same.
582
583 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
584 as not to get warnings when autogen.sh is called. Also changed
585 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
586
587 44. To help people who use pcregrep to scan files containing exceedingly long
588 lines, the following changes have been made:
589
590 (a) The default value of the buffer size parameter has been increased from
591 8K to 20K. (The actual buffer used is three times this size.)
592
593 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
594 PCRE is built.
595
596 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
597 to be set at run time.
598
599 (d) Numerical values in pcregrep options can be followed by K or M, for
600 example --buffer-size=50K.
601
602 (e) If a line being scanned overflows pcregrep's buffer, an error is now
603 given and the return code is set to 2.
604
605 45. Add a pointer to the latest mark to the callout data block.
606
607 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
608 partial match of an empty string instead of no match. This was specific to
609 the use of ".".
610
611 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
612 complete match instead of a partial match. This bug was dependent on both
613 the PCRE_UTF8 and PCRE_DOTALL options being set.
614
615 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
616 starting byte set, because \b was not being ignored.
617
618
619 Version 8.12 15-Jan-2011
620 ------------------------
621
622 1. Fixed some typos in the markup of the man pages, and wrote a script that
623 checks for such things as part of the documentation building process.
624
625 2. On a big-endian 64-bit system, pcregrep did not correctly process the
626 --match-limit and --recursion-limit options (added for 8.11). In
627 particular, this made one of the standard tests fail. (The integer value
628 went into the wrong half of a long int.)
629
630 3. If the --colour option was given to pcregrep with -v (invert match), it
631 did strange things, either producing crazy output, or crashing. It should,
632 of course, ignore a request for colour when reporting lines that do not
633 match.
634
635 4. Another pcregrep bug caused similar problems if --colour was specified with
636 -M (multiline) and the pattern match finished with a line ending.
637
638 5. In pcregrep, when a pattern that ended with a literal newline sequence was
639 matched in multiline mode, the following line was shown as part of the
640 match. This seems wrong, so I have changed it.
641
642 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
643 the check for further matches in the same line (so they could be coloured)
644 to overrun the end of the current line. If another match was found, it was
645 incorrectly shown (and then shown again when found in the next line).
646
647 7. If pcregrep was compiled under Windows, there was a reference to the
648 function pcregrep_exit() before it was defined. I am assuming this was
649 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
650 reported by a user. I've moved the definition above the reference.
651
652
653 Version 8.11 10-Dec-2010
654 ------------------------
655
656 1. (*THEN) was not working properly if there were untried alternatives prior
657 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
658 backtracked to try for "b" instead of moving to the next alternative branch
659 at the same level (in this case, to look for "c"). The Perl documentation
660 is clear that when (*THEN) is backtracked onto, it goes to the "next
661 alternative in the innermost enclosing group".
662
663 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
664 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
665 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
666 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
667 (*THEN).
668
669 3. If \s appeared in a character class, it removed the VT character from
670 the class, even if it had been included by some previous item, for example
671 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
672 of \s, but is part of the POSIX "space" class.)
673
674 4. A partial match never returns an empty string (because you can always
675 match an empty string at the end of the subject); however the checking for
676 an empty string was starting at the "start of match" point. This has been
677 changed to the "earliest inspected character" point, because the returned
678 data for a partial match starts at this character. This means that, for
679 example, /(?<=abc)def/ gives a partial match for the subject "abc"
680 (previously it gave "no match").
681
682 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
683 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
684 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
685 has an implication that the given string is incomplete (because a partial
686 match is preferred over a full match). For this reason, these items now
687 give a partial match in this situation. [Aside: previously, the one case
688 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
689 match rather than a full match, which was wrong by the old rules, but is
690 now correct.]
691
692 6. There was a bug in the handling of #-introduced comments, recognized when
693 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
694 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
695 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
696 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
697 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
698 places in pcre_compile().
699
700 7. Related to (6) above, when pcre_compile() was skipping #-introduced
701 comments when looking ahead for named forward references to subpatterns,
702 the only newline sequence it recognized was NL. It now handles newlines
703 according to the set newline convention.
704
705 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
706 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
707 cater for a lack of strerror(). These oversights have been fixed.
708
709 9. Added --match-limit and --recursion-limit to pcregrep.
710
711 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
712
713 11. When the -o option was used, pcregrep was setting a return code of 1, even
714 when matches were found, and --line-buffered was not being honoured.
715
716 12. Added an optional parentheses number to the -o and --only-matching options
717 of pcregrep.
718
719 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
720 can match an empty string. The code to do it in pcretest and pcredemo
721 needed fixing:
722
723 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
724 only one byte after an empty string match just before CRLF (this case
725 just got forgotten; "any" and "anycrlf" were OK).
726
727 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
728 mode when an empty string match preceded an ASCII character followed by
729 a non-ASCII character. (The code for advancing by one character rather
730 than one byte was nonsense.)
731
732 (c) The pcredemo.c sample program did not have any code at all to handle
733 the cases when CRLF is a valid newline sequence.
734
735 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
736 as a starting offset was within the subject string. There is now a new
737 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
738 negative or greater than the length of the string. In order to test this,
739 pcretest is extended to allow the setting of negative starting offsets.
740
741 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
742 starting offset points to the beginning of a UTF-8 character was
743 unnecessarily clumsy. I tidied it up.
744
745 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
746 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
747
748 17. Nobody had reported that the --include_dir option, which was added in
749 release 7.7 should have been called --include-dir (hyphen, not underscore)
750 for compatibility with GNU grep. I have changed it to --include-dir, but
751 left --include_dir as an undocumented synonym, and the same for
752 --exclude-dir, though that is not available in GNU grep, at least as of
753 release 2.5.4.
754
755 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
756 characters from a string of bytes) have been redefined so as not to use
757 loops, in order to improve performance in some environments. At the same
758 time, I abstracted some of the common code into auxiliary macros to save
759 repetition (this should not affect the compiled code).
760
761 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
762 compile-time error is now given if \c is not followed by an ASCII
763 character, that is, a byte less than 128. (In EBCDIC mode, the code is
764 different, and any byte value is allowed.)
765
766 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
767 START_OPTIMIZE option, which is now allowed at compile time - but just
768 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
769 to pcregrep and other applications that have no direct access to PCRE
770 options. The new /Y option in pcretest sets this option when calling
771 pcre_compile().
772
773 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
774 back references. Groups containing recursive back references were forced to
775 be atomic by that change, but in the case of named groups, the amount of
776 memory required was incorrectly computed, leading to "Failed: internal
777 error: code overflow". This has been fixed.
778
779 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
780 pcretest.c, to avoid build problems in some Borland environments.
781
782
783 Version 8.10 25-Jun-2010
784 ------------------------
785
786 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
787 THEN.
788
789 2. (*ACCEPT) was not working when inside an atomic group.
790
791 3. Inside a character class, \B is treated as a literal by default, but
792 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
793 causes the error). The code is unchanged, but I tidied the documentation.
794
795 4. Inside a character class, PCRE always treated \R and \X as literals,
796 whereas Perl faults them if its -w option is set. I have changed PCRE so
797 that it faults them when PCRE_EXTRA is set.
798
799 5. Added support for \N, which always matches any character other than
800 newline. (It is the same as "." when PCRE_DOTALL is not set.)
801
802 6. When compiling pcregrep with newer versions of gcc which may have
803 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
804 declared with attribute warn_unused_result" were given. Just casting the
805 result to (void) does not stop the warnings; a more elaborate fudge is
806 needed. I've used a macro to implement this.
807
808 7. Minor change to pcretest.c to avoid a compiler warning.
809
810 8. Added four artifical Unicode properties to help with an option to make
811 \s etc use properties (see next item). The new properties are: Xan
812 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
813
814 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
815 use Unicode properties. (*UCP) at the start of a pattern can be used to set
816 this option. Modified pcretest to add /W to test this facility. Added
817 REG_UCP to make it available via the POSIX interface.
818
819 10. Added --line-buffered to pcregrep.
820
821 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
822 studied, and the match started with a letter with a code point greater than
823 127 whose first byte was different to the first byte of the other case of
824 the letter, the other case of this starting letter was not recognized
825 (#976).
826
827 12. If a pattern that was studied started with a repeated Unicode property
828 test, for example, \p{Nd}+, there was the theoretical possibility of
829 setting up an incorrect bitmap of starting bytes, but fortunately it could
830 not have actually happened in practice until change 8 above was made (it
831 added property types that matched character-matching opcodes).
832
833 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
834 possible starting bytes for non-anchored patterns.
835
836 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
837 \R, and also a number of cases that involve Unicode properties, both
838 explicit and implicit when PCRE_UCP is set.
839
840 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
841 input, it could crash or give wrong results if characters with values
842 greater than 0xc0 were present in the subject string. (Detail: it assumed
843 UTF-8 input when processing these items.)
844
845 16. Added a lot of (int) casts to avoid compiler warnings in systems where
846 size_t is 64-bit (#991).
847
848 17. Added a check for running out of memory when PCRE is compiled with
849 --disable-stack-for-recursion (#990).
850
851 18. If the last data line in a file for pcretest does not have a newline on
852 the end, a newline was missing in the output.
853
854 19. The default pcre_chartables.c file recognizes only ASCII characters (values
855 less than 128) in its various bitmaps. However, there is a facility for
856 generating tables according to the current locale when PCRE is compiled. It
857 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
858 characters, are recognized by isspace() and therefore were getting set in
859 these tables, and indeed these tables seem to approximate to ISO 8859. This
860 caused a problem in UTF-8 mode when pcre_study() was used to create a list
861 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
862 which of course cannot start UTF-8 characters. I have changed the code so
863 that only real ASCII characters (less than 128) and the correct starting
864 bytes for UTF-8 encodings are set for characters greater than 127 when in
865 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
866 altogether.)
867
868 20. Added the /T option to pcretest so as to be able to run tests with non-
869 standard character tables, thus making it possible to include the tests
870 used for 19 above in the standard set of tests.
871
872 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
873 reference to a subpattern the other side of a comment that contains an
874 opening parenthesis caused either an internal compiling error, or a
875 reference to the wrong subpattern.
876
877
878 Version 8.02 19-Mar-2010
879 ------------------------
880
881 1. The Unicode data tables have been updated to Unicode 5.2.0.
882
883 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
884 configured.
885
886 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
887 original author of that file, following a query about its status.
888
889 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
890 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
891
892 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
893 quantifier applied to a forward-referencing subroutine call, could compile
894 incorrect code or give the error "internal error: previously-checked
895 referenced subpattern not found".
896
897 6. Both MS Visual Studio and Symbian OS have problems with initializing
898 variables to point to external functions. For these systems, therefore,
899 pcre_malloc etc. are now initialized to local functions that call the
900 relevant global functions.
901
902 7. There were two entries missing in the vectors called coptable and poptable
903 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
904 I've fixed the data, and added a kludgy way of testing at compile time that
905 the lengths are correct (equal to the number of opcodes).
906
907 8. Following on from 7, I added a similar kludge to check the length of the
908 eint vector in pcreposix.c.
909
910 9. Error texts for pcre_compile() are held as one long string to avoid too
911 much relocation at load time. To find a text, the string is searched,
912 counting zeros. There was no check for running off the end of the string,
913 which could happen if a new error number was added without updating the
914 string.
915
916 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
917
918 11. \K was not working if it appeared in an atomic group or in a group that
919 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
920 \K is "not well defined" if used in an assertion. PCRE now accepts it if
921 the assertion is positive, but not if it is negative.
922
923 12. Change 11 fortuitously reduced the size of the stack frame used in the
924 "match()" function of pcre_exec.c by one pointer. Forthcoming
925 implementation of support for (*MARK) will need an extra pointer on the
926 stack; I have reserved it now, so that the stack frame size does not
927 decrease.
928
929 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
930 item in branch that calls a recursion is a subroutine call - as in the
931 second branch in the above example - was incorrectly given the compile-
932 time error "recursive call could loop indefinitely" because pcre_compile()
933 was not correctly checking the subroutine for matching a non-empty string.
934
935 14. The checks for overrunning compiling workspace could trigger after an
936 overrun had occurred. This is a "should never occur" error, but it can be
937 triggered by pathological patterns such as hundreds of nested parentheses.
938 The checks now trigger 100 bytes before the end of the workspace.
939
940 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
941
942
943 Version 8.01 19-Jan-2010
944 ------------------------
945
946 1. If a pattern contained a conditional subpattern with only one branch (in
947 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
948 computed the wrong minimum data length (which is of course zero for such
949 subpatterns). This could cause incorrect "no match" results.
950
951 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
952 the pattern is reset in the first branch, pcre_compile() failed with
953 "internal error: code overflow at offset...". This happened only when
954 the reset was to the original external option setting. (An optimization
955 abstracts leading options settings into an external setting, which was the
956 cause of this.)
957
958 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
959 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
960 assertion pattern did not match (meaning that the assertion was true), it
961 was incorrectly treated as false if the SKIP had been reached during the
962 matching. This also applied to assertions used as conditions.
963
964 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
965 assertion subpattern, including such a pattern used as a condition,
966 unpredictable results occurred, instead of the error return
967 PCRE_ERROR_DFA_UITEM.
968
969 5. The C++ GlobalReplace function was not working like Perl for the special
970 situation when an empty string is matched. It now does the fancy magic
971 stuff that is necessary.
972
973 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
974 removed. (These were left over from very, very early versions of PCRE.)
975
976 7. Some cosmetic changes to the code to make life easier when compiling it
977 as part of something else:
978
979 (a) Change DEBUG to PCRE_DEBUG.
980
981 (b) In pcre_compile(), rename the member of the "branch_chain" structure
982 called "current" as "current_branch", to prevent a collision with the
983 Linux macro when compiled as a kernel module.
984
985 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
986 prevent a collision with the Linux macro when compiled as a kernel
987 module.
988
989 8. In pcre_compile() there are some checks for integer overflows that used to
990 cast potentially large values to (double). This has been changed to that
991 when building, a check for int64_t is made, and if it is found, it is used
992 instead, thus avoiding the use of floating point arithmetic. (There is no
993 other use of FP in PCRE.) If int64_t is not found, the fallback is to
994 double.
995
996 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
997 2005 (difference between two addresses compared to an unsigned value).
998
999 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1000 custom one, because of the following reported problem in Windows:
1001
1002 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1003 under Win32.
1004 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1005 therefore missing the function definition.
1006 - The compiler thus generates a "C" signature for the test function.
1007 - The linker fails to find the "C" function.
1008 - PCRE fails to configure if asked to do so against libbz2.
1009
1010 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1011 messages were output:
1012
1013 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1014 rerunning libtoolize, to keep the correct libtool macros in-tree.
1015 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1016
1017 I have done both of these things.
1018
1019 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1020 most of the time, it *can* run out if it is given a pattern that contains a
1021 runaway infinite recursion. I updated the discussion in the pcrestack man
1022 page.
1023
1024 13. Now that we have gone to the x.xx style of version numbers, the minor
1025 version may start with zero. Using 08 or 09 is a bad idea because users
1026 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1027 interpreted as invalid octal numbers. I've updated the previous comment in
1028 configure.ac, and also added a check that gives an error if 08 or 09 are
1029 used.
1030
1031 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1032 causing partial matching to fail when the end of the subject matched \W
1033 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1034
1035 15. There were some discrepancies between the declarations in pcre_internal.h
1036 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1037 their definitions. The declarations used "const uschar *" and the
1038 definitions used USPTR. Even though USPTR is normally defined as "const
1039 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1040 reported that: "This difference in casting confuses some C++ compilers, for
1041 example, SunCC recognizes above declarations as different functions and
1042 generates broken code for hbpcre." I have changed the declarations to use
1043 USPTR.
1044
1045 16. GNU libtool is named differently on some systems. The autogen.sh script now
1046 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1047 (FreeBSD).
1048
1049 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1050 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1051 comment: "Figure out how to create a longlong from a string: strtoll and
1052 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1053 instance, but it only takes 2 args instead of 3!"
1054
1055 18. A subtle bug concerned with back references has been fixed by a change of
1056 specification, with a corresponding code fix. A pattern such as
1057 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1058 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1059 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1060 same bug. Such groups have to be quantified to be useful, or contained
1061 inside another quantified group. (If there's no repetition, the reference
1062 can never match.) The problem arises because, having left the group and
1063 moved on to the rest of the pattern, a later failure that backtracks into
1064 the group uses the captured value from the final iteration of the group
1065 rather than the correct earlier one. I have fixed this in PCRE by forcing
1066 any group that contains a reference to itself to be an atomic group; that
1067 is, there cannot be any backtracking into it once it has completed. This is
1068 similar to recursive and subroutine calls.
1069
1070
1071 Version 8.00 19-Oct-09
1072 ----------------------
1073
1074 1. The table for translating pcre_compile() error codes into POSIX error codes
1075 was out-of-date, and there was no check on the pcre_compile() error code
1076 being within the table. This could lead to an OK return being given in
1077 error.
1078
1079 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1080 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1081 in a Windows environment.
1082
1083 3. The pcregrep --count option prints the count for each file even when it is
1084 zero, as does GNU grep. However, pcregrep was also printing all files when
1085 --files-with-matches was added. Now, when both options are given, it prints
1086 counts only for those files that have at least one match. (GNU grep just
1087 prints the file name in this circumstance, but including the count seems
1088 more useful - otherwise, why use --count?) Also ensured that the
1089 combination -clh just lists non-zero counts, with no names.
1090
1091 4. The long form of the pcregrep -F option was incorrectly implemented as
1092 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1093 but it seems right to fix it, and I didn't think it was worth preserving
1094 the old behaviour.
1095
1096 5. The command line items --regex=pattern and --regexp=pattern were not
1097 recognized by pcregrep, which required --regex pattern or --regexp pattern
1098 (with a space rather than an '='). The man page documented the '=' forms,
1099 which are compatible with GNU grep; these now work.
1100
1101 6. No libpcreposix.pc file was created for pkg-config; there was just
1102 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1103
1104 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1105 when UCP support is not needed, by modifying the Python script that
1106 generates it from Unicode data files. This should not matter if the module
1107 is correctly used as a library, but I received one complaint about 50K of
1108 unwanted data. My guess is that the person linked everything into his
1109 program rather than using a library. Anyway, it does no harm.
1110
1111 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1112 was a minimum greater than 1 for a wide character in a possessive
1113 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1114 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1115 character. Chaos in the form of incorrect output or a compiling loop could
1116 result.
1117
1118 9. The restrictions on what a pattern can contain when partial matching is
1119 requested for pcre_exec() have been removed. All patterns can now be
1120 partially matched by this function. In addition, if there are at least two
1121 slots in the offset vector, the offset of the earliest inspected character
1122 for the match and the offset of the end of the subject are set in them when
1123 PCRE_ERROR_PARTIAL is returned.
1124
1125 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1126 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1127 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1128 and may be more useful for multi-segment matching.
1129
1130 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1131 used to be given if ever the end of the subject was reached; now it is
1132 given only if matching could not proceed because another character was
1133 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1134 string "Z", which now yields "no match" instead of "partial match". In the
1135 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1136 final character ended with (*FAIL).
1137
1138 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1139 if the pattern had a "must contain" character that was already found in the
1140 earlier partial match, unless partial matching was again requested. For
1141 example, with the pattern /dog.(body)?/, the "must contain" character is
1142 "g". If the first part-match was for the string "dog", restarting with
1143 "sbody" failed. This bug has been fixed.
1144
1145 13. The string returned by pcre_dfa_exec() after a partial match has been
1146 changed so that it starts at the first inspected character rather than the
1147 first character of the match. This makes a difference only if the pattern
1148 starts with a lookbehind assertion or \b or \B (\K is not supported by
1149 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1150 matching functions compatible, and I think it's the right thing to do.
1151
1152 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1153 so that the demonstration program is easily available in environments where
1154 PCRE has not been installed from source.
1155
1156 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1157 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1158 library.
1159
1160 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1161 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1162 is not the first non-POSIX option to be added. Clearly some people find
1163 these options useful.
1164
1165 17. If a caller to the POSIX matching function regexec() passes a non-zero
1166 value for nmatch with a NULL value for pmatch, the value of
1167 nmatch is forced to zero.
1168
1169 18. RunGrepTest did not have a test for the availability of the -u option of
1170 the diff command, as RunTest does. It now checks in the same way as
1171 RunTest, and also checks for the -b option.
1172
1173 19. If an odd number of negated classes containing just a single character
1174 interposed, within parentheses, between a forward reference to a named
1175 subpattern and the definition of the subpattern, compilation crashed with
1176 an internal error, complaining that it could not find the referenced
1177 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1178 [The bug was that it was starting one character too far in when skipping
1179 over the character class, thus treating the ] as data rather than
1180 terminating the class. This meant it could skip too much.]
1181
1182 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1183 /g option in pcretest when the pattern contains \K, which makes it possible
1184 to have an empty string match not at the start, even when the pattern is
1185 anchored. Updated pcretest and pcredemo to use this option.
1186
1187 21. If the maximum number of capturing subpatterns in a recursion was greater
1188 than the maximum at the outer level, the higher number was returned, but
1189 with unset values at the outer level. The correct (outer level) value is
1190 now given.
1191
1192 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1193 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1194 make it do so. The string so far is captured, making this feature
1195 compatible with Perl.
1196
1197 23. The tests have been re-organized, adding tests 11 and 12, to make it
1198 possible to check the Perl 5.10 features against Perl 5.10.
1199
1200 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1201 pattern matches a fixed length string. PCRE did not allow this; now it
1202 does. Neither allows recursion.
1203
1204 25. I finally figured out how to implement a request to provide the minimum
1205 length of subject string that was needed in order to match a given pattern.
1206 (It was back references and recursion that I had previously got hung up
1207 on.) This code has now been added to pcre_study(); it finds a lower bound
1208 to the length of subject needed. It is not necessarily the greatest lower
1209 bound, but using it to avoid searching strings that are too short does give
1210 some useful speed-ups. The value is available to calling programs via
1211 pcre_fullinfo().
1212
1213 26. While implementing 25, I discovered to my embarrassment that pcretest had
1214 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1215 study optimizations had never been tested with that matching function.
1216 Oops. What is worse, even when it was passed study data, there was a bug in
1217 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1218 were also very few tests of studied patterns with pcre_dfa_exec().
1219
1220 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1221 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1222 on the other side of the coin, they are no longer allowed to have different
1223 names, because these cannot be distinguished in PCRE, and this has caused
1224 confusion. (This is a difference from Perl.)
1225
1226 28. When duplicate subpattern names are present (necessarily with different
1227 numbers, as required by 27 above), and a test is made by name in a
1228 conditional pattern, either for a subpattern having been matched, or for
1229 recursion in such a pattern, all the associated numbered subpatterns are
1230 tested, and the overall condition is true if the condition is true for any
1231 one of them. This is the way Perl works, and is also more like the way
1232 testing by number works.
1233
1234
1235 Version 7.9 11-Apr-09
1236 ---------------------
1237
1238 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1239 (pcretest), all targets were linked against these libraries. This included
1240 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1241 libraries. This caused unwanted dependencies to be created. This problem
1242 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1243 pcretest is linked with readline.
1244
1245 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1246 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1247 moved outside it again, because FALSE and TRUE are already defined in AIX,
1248 but BOOL is not.
1249
1250 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1251 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1252
1253 4. The pcregrep documentation said spaces were inserted as well as colons (or
1254 hyphens) following file names and line numbers when outputting matching
1255 lines. This is not true; no spaces are inserted. I have also clarified the
1256 wording for the --colour (or --color) option.
1257
1258 5. In pcregrep, when --colour was used with -o, the list of matching strings
1259 was not coloured; this is different to GNU grep, so I have changed it to be
1260 the same.
1261
1262 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1263 each matching line was coloured. Now it goes on to look for further matches
1264 of any of the test patterns, which is the same behaviour as GNU grep.
1265
1266 7. A pattern that could match an empty string could cause pcregrep to loop; it
1267 doesn't make sense to accept an empty string match in pcregrep, so I have
1268 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1269 seems to be how GNU grep behaves.
1270
1271 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1272 start or after a newline", because the conditional assertion was not being
1273 correctly handled. The rule now is that both the assertion and what follows
1274 in the first alternative must satisfy the test.
1275
1276 9. If auto-callout was enabled in a pattern with a conditional group whose
1277 condition was an assertion, PCRE could crash during matching, both with
1278 pcre_exec() and pcre_dfa_exec().
1279
1280 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1281 used for matching.
1282
1283 11. Unicode property support in character classes was not working for
1284 characters (bytes) greater than 127 when not in UTF-8 mode.
1285
1286 12. Added the -M command line option to pcretest.
1287
1288 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1289
1290 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1291
1292 16. Added comments and documentation about mis-use of no_arg in the C++
1293 wrapper.
1294
1295 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1296 from Martin Jerabek that uses macro names for all relevant character and
1297 string constants.
1298
1299 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1300 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1301 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1302 these, but not everybody uses configure.
1303
1304 19. A conditional group that had only one branch was not being correctly
1305 recognized as an item that could match an empty string. This meant that an
1306 enclosing group might also not be so recognized, causing infinite looping
1307 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1308 with the subject "ab", where knowledge that the repeated group can match
1309 nothing is needed in order to break the loop.
1310
1311 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1312 exec(), but without supplying a callout function, matching went wrong.
1313
1314 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1315 leak if the size of the offset vector was greater than 30. When the vector
1316 is smaller, the saved offsets during recursion go onto a local stack
1317 vector, but for larger vectors malloc() is used. It was failing to free
1318 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1319 error, in fact).
1320
1321 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1322 heapframe that is used only when UTF-8 support is enabled. This caused no
1323 problem, but was untidy.
1324
1325 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1326 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1327 included within another project.
1328
1329 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1330 slightly modified by me:
1331
1332 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1333 not building pcregrep.
1334
1335 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1336 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1337
1338 25. Forward references, both numeric and by name, in patterns that made use of
1339 duplicate group numbers, could behave incorrectly or give incorrect errors,
1340 because when scanning forward to find the reference group, PCRE was not
1341 taking into account the duplicate group numbers. A pattern such as
1342 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1343
1344 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1345 the feature of a custom pointer more persuasive (as requested by a user).
1346
1347 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1348 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1349 pre-defined.
1350
1351 28. Added support for (*UTF8) at the start of a pattern.
1352
1353 29. Arrange for flags added by the "release type" setting in CMake to be shown
1354 in the configuration summary.
1355
1356
1357 Version 7.8 05-Sep-08
1358 ---------------------
1359
1360 1. Replaced UCP searching code with optimized version as implemented for Ad
1361 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1362 stage table and inline lookup instead of a function, giving speed ups of 2
1363 to 5 times on some simple patterns that I tested. Permission was given to
1364 distribute the MultiStage2.py script that generates the tables (it's not in
1365 the tarball, but is in the Subversion repository).
1366
1367 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1368 scripts.
1369
1370 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1371 a group with a zero qualifier. The result of the study could be incorrect,
1372 or the function might crash, depending on the pattern.
1373
1374 4. Caseless matching was not working for non-ASCII characters in back
1375 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1376 It now works when Unicode Property Support is available.
1377
1378 5. In pcretest, an escape such as \x{de} in the data was always generating
1379 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1380 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1381 truncation.
1382
1383 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1384
1385 7. Added two (int) casts to pcregrep when printing the difference of two
1386 pointers, in case they are 64-bit values.
1387
1388 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1389 test 2 if it fails.
1390
1391 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1392 and a #define of that name to empty if it is not externally set. This is to
1393 allow users of MSVC to set it if necessary.
1394
1395 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1396 the convenience functions in the pcre_get.c source file.
1397
1398 11. An option change at the start of a pattern that had top-level alternatives
1399 could cause overwriting and/or a crash. This command provoked a crash in
1400 some environments:
1401
1402 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1403
1404 This potential security problem was recorded as CVE-2008-2371.
1405
1406 12. For a pattern where the match had to start at the beginning or immediately
1407 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1408 pcre_dfa_exec() could read past the end of the passed subject if there was
1409 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1410 pcretest so that it places the subject at the end of its malloc-ed buffer.
1411
1412 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1413 exec() might read past the end of the data buffer in UTF-8 mode.
1414
1415 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1416 the data contained the byte 0x85 as part of a UTF-8 character within its
1417 first line. This applied both to normal and DFA matching.
1418
1419 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1420 /^[^d]*?$/8 failed to match "abc".
1421
1422 16. Added a missing copyright notice to pcrecpp_internal.h.
1423
1424 17. Make it more clear in the documentation that values returned from
1425 pcre_exec() in ovector are byte offsets, not character counts.
1426
1427 18. Tidied a few places to stop certain compilers from issuing warnings.
1428
1429 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1430 supplied by Stefan Weber. I made a further small update for 7.8 because
1431 there is a change of source arrangements: the pcre_searchfuncs.c module is
1432 replaced by pcre_ucd.c.
1433
1434
1435 Version 7.7 07-May-08
1436 ---------------------
1437
1438 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1439 a string to a long long, pretend we don't even have a long long." This is
1440 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1441
1442 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1443 pre-7.6 versions, which defined a global no_arg variable instead of putting
1444 it in the RE class. (See also #8 below.)
1445
1446 3. Remove a line of dead code, identified by coverity and reported by Nuno
1447 Lopes.
1448
1449 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1450
1451 (1) The include/exclude patterns were being applied to the whole pathnames
1452 of files, instead of just to the final components.
1453
1454 (2) If there was more than one level of directory, the subdirectories were
1455 skipped unless they satisfied the include/exclude conditions. This is
1456 inconsistent with GNU grep (and could even be seen as contrary to the
1457 pcregrep specification - which I improved to make it absolutely clear).
1458 The action now is always to scan all levels of directory, and just
1459 apply the include/exclude patterns to regular files.
1460
1461 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1462 --exclude_dir in the tests to avoid scanning .svn directories.
1463
1464 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1465 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1466 doesn't support NULs in patterns.
1467
1468 7. Added some missing "const"s to declarations of static tables in
1469 pcre_compile.c and pcre_dfa_exec.c.
1470
1471 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1472 caused by fix #2 above. (Subsequently also a second patch to fix the
1473 first patch. And a third patch - this was a messy problem.)
1474
1475 9. Applied Craig's patch to remove the use of push_back().
1476
1477 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1478 matching function regexec().
1479
1480 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1481 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1482 references. PCRE supports relative numbers with this syntax (I don't think
1483 Oniguruma does).
1484
1485 12. Previously, a group with a zero repeat such as (...){0} was completely
1486 omitted from the compiled regex. However, this means that if the group
1487 was called as a subroutine from elsewhere in the pattern, things went wrong
1488 (an internal error was given). Such groups are now left in the compiled
1489 pattern, with a new opcode that causes them to be skipped at execution
1490 time.
1491
1492 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1493 to the way PCRE behaves:
1494
1495 (a) A lone ] character is dis-allowed (Perl treats it as data).
1496
1497 (b) A back reference to an unmatched subpattern matches an empty string
1498 (Perl fails the current match path).
1499
1500 (c) A data ] in a character class must be notated as \] because if the
1501 first data character in a class is ], it defines an empty class. (In
1502 Perl it is not possible to have an empty class.) The empty class []
1503 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1504 The negative empty class [^] matches any one character, independently
1505 of the DOTALL setting.
1506
1507 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1508 non-existent subpattern following a character class starting with ']' and
1509 containing () gave an internal compiling error instead of "reference to
1510 non-existent subpattern". Fortunately, when the pattern did exist, the
1511 compiled code was correct. (When scanning forwards to check for the
1512 existencd of the subpattern, it was treating the data ']' as terminating
1513 the class, so got the count wrong. When actually compiling, the reference
1514 was subsequently set up correctly.)
1515
1516 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1517 it was being rejected as not supported by pcre_dfa_exec(), even though
1518 other assertions are supported. I have made pcre_dfa_exec() support
1519 (*FAIL).
1520
1521 16. The implementation of 13c above involved the invention of a new opcode,
1522 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1523 cannot be changed at match time, I realized I could make a small
1524 improvement to matching performance by compiling OP_ALLANY instead of
1525 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1526 on the OP_ANY path.
1527
1528 17. Compiling pcretest on Windows with readline support failed without the
1529 following two fixes: (1) Make the unistd.h include conditional on
1530 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1531
1532 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1533 ncurses library to be included for pcretest when ReadLine support is
1534 requested, but also to allow for it to be overridden. This patch came from
1535 Daniel Bergström.
1536
1537 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1538 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1539 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1540 spotting this.
1541
1542
1543 Version 7.6 28-Jan-08
1544 ---------------------
1545
1546 1. A character class containing a very large number of characters with
1547 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1548 overflow.
1549
1550 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1551 HAVE_LONG_LONG is not defined.
1552
1553 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1554 bring them up to date and include new features. This patch includes:
1555
1556 - Fixed PH's badly added libz and libbz2 support.
1557 - Fixed a problem with static linking.
1558 - Added pcredemo. [But later removed - see 7 below.]
1559 - Fixed dftables problem and added an option.
1560 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1561 HAVE_LONG_LONG.
1562 - Added readline support for pcretest.
1563 - Added an listing of the option settings after cmake has run.
1564
1565 4. A user submitted a patch to Makefile that makes it easy to create
1566 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1567 Makefile.am that cause it to include this special target, without
1568 affecting anything else. Note that the same mingw target plus all
1569 the other distribution libraries and programs are now supported
1570 when configuring with CMake (see 6 below) instead of with
1571 Configure/Make.
1572
1573 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1574 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1575 exported in the Windows port". It has not yet been confirmed that the patch
1576 solves the problem, but it does no harm.
1577
1578 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1579 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1580 with CMake, and also correct the comment about stack recursion.
1581
1582 7. Remove the automatic building of pcredemo from the ./configure system and
1583 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1584 of a program that users should build themselves after PCRE is installed, so
1585 building it automatically is not really right. What is more, it gave
1586 trouble in some build environments.
1587
1588 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1589
1590
1591 Version 7.5 10-Jan-08
1592 ---------------------
1593
1594 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1595 values in parens when parsing an RE using the C++ wrapper."
1596
1597 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1598 Characters greater than 255 were excluded from the class instead of being
1599 included.
1600
1601 3. The same bug as (2) above applied to negated POSIX classes such as
1602 [:^space:].
1603
1604 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1605 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1606 I have changed it.
1607
1608 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1609 first named subpattern) and a construct such as (?&a) would reference the
1610 first named subpattern whose name started with "a" (in other words, the
1611 length check was missing). Both these problems are fixed. "Subpattern name
1612 expected" is now given for (?&) (a zero-length name), and this patch also
1613 makes it give the same error for \k'' (previously it complained that that
1614 was a reference to a non-existent subpattern).
1615
1616 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1617 this is right because (?- can be followed by option settings as well as by
1618 digits. I have, however, made the messages clearer.
1619
1620 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1621 than the number used in the conditional) now cause a compile-time error.
1622 This is actually not compatible with Perl, which accepts such patterns, but
1623 treats the conditional as always being FALSE (as PCRE used to), but it
1624 seems to me that giving a diagnostic is better.
1625
1626 8. Change "alphameric" to the more common word "alphanumeric" in comments
1627 and messages.
1628
1629 9. Fix two occurrences of "backslash" in comments that should have been
1630 "backspace".
1631
1632 10. Remove two redundant lines of code that can never be obeyed (their function
1633 was moved elsewhere).
1634
1635 11. The program that makes PCRE's Unicode character property table had a bug
1636 which caused it to generate incorrect table entries for sequences of
1637 characters that have the same character type, but are in different scripts.
1638 It amalgamated them into a single range, with the script of the first of
1639 them. In other words, some characters were in the wrong script. There were
1640 thirteen such cases, affecting characters in the following ranges:
1641
1642 U+002b0 - U+002c1
1643 U+0060c - U+0060d
1644 U+0061e - U+00612
1645 U+0064b - U+0065e
1646 U+0074d - U+0076d
1647 U+01800 - U+01805
1648 U+01d00 - U+01d77
1649 U+01d9b - U+01dbf
1650 U+0200b - U+0200f
1651 U+030fc - U+030fe
1652 U+03260 - U+0327f
1653 U+0fb46 - U+0fbb1
1654 U+10450 - U+1049d
1655
1656 12. The -o option (show only the matching part of a line) for pcregrep was not
1657 compatible with GNU grep in that, if there was more than one match in a
1658 line, it showed only the first of them. It now behaves in the same way as
1659 GNU grep.
1660
1661 13. If the -o and -v options were combined for pcregrep, it printed a blank
1662 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1663 does the same. The return code can be used to tell if there were any
1664 non-matching lines.
1665
1666 14. Added --file-offsets and --line-offsets to pcregrep.
1667
1668 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1669 infinitely looping recursion. The bug was that positive lookaheads were not
1670 being skipped when checking for a possible empty match (negative lookaheads
1671 and both kinds of lookbehind were skipped).
1672
1673 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1674 inclusion of <windows.h> to before rather than after the definition of
1675 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1676
1677 17. Specifying a possessive quantifier with a specific limit for a Unicode
1678 character property caused pcre_compile() to compile bad code, which led at
1679 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1680 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1681 caused the error; without that there was no problem.
1682
1683 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1684
1685 19. Added --enable-pcretest-libreadline.
1686
1687 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1688 RE::GlobalReplace(). As a result, the number of replacements returned was
1689 double what it should be. I removed one of the increments, but Craig sent a
1690 later patch that removed the other one (the right fix) and added unit tests
1691 that check the return values (which was not done before).
1692
1693 21. Several CMake things:
1694
1695 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1696 the names libpcre and libpcreposix, not just pcre and pcreposix.
1697
1698 (2) The above change means that pcretest and pcregrep are now correctly
1699 linked with the newly-built libraries, not previously installed ones.
1700
1701 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1702
1703 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1704 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1705 UTF-8 newline character). The key issue is that the pattern starts .*;
1706 this means that the match must be either at the beginning, or after a
1707 newline. The bug was in the code for advancing after a failed match and
1708 checking that the new position followed a newline. It was not taking
1709 account of UTF-8 characters correctly.
1710
1711 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1712 character classes. PCRE was not treating the sequence [:...:] as a
1713 character class unless the ... were all letters. Perl, however, seems to
1714 allow any characters between [: and :], though of course it rejects as
1715 unknown any "names" that contain non-letters, because all the known class
1716 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1717 for example, whereas PCRE did not - it did not recognize a POSIX character
1718 class. This seemed a bit dangerous, so the code has been changed to be
1719 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1720 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1721 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1722 Perl does, and where it didn't before.
1723
1724 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1725 Windows environments %n is disabled by default.
1726
1727
1728 Version 7.4 21-Sep-07
1729 ---------------------
1730
1731 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1732 means that a class such as [\s] counted as "explicit reference to CR or
1733 LF". That isn't really right - the whole point of the change was to try to
1734 help when there was an actual mention of one of the two characters. So now
1735 the change happens only if \r or \n (or a literal CR or LF) character is
1736 encountered.
1737
1738 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1739 of both had grown to the point where there were only 3 bits left.
1740 Fortunately, there was spare space in the data structure, and so I have
1741 moved the internal flags into a new 16-bit field to free up more option
1742 bits.
1743
1744 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1745 but did not set the internal JCHANGED flag - either of these is enough to
1746 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1747 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1748 start sets both bits.
1749
1750 4. Added options (at build time, compile time, exec time) to change \R from
1751 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1752
1753 5. doc/pcresyntax.html was missing from the distribution.
1754
1755 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1756 compatibility, even though it is no longer used.
1757
1758 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1759 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1760 windows.h file is present (where different names are used). [This was
1761 reversed later after testing - see 16 below.]
1762
1763 8. Changed all #include <config.h> to #include "config.h". There were also
1764 some further <pcre.h> cases that I changed to "pcre.h".
1765
1766 9. When pcregrep was used with the --colour option, it missed the line ending
1767 sequence off the lines that it output.
1768
1769 10. It was pointed out to me that arrays of string pointers cause lots of
1770 relocations when a shared library is dynamically loaded. A technique of
1771 using a single long string with a table of offsets can drastically reduce
1772 these. I have refactored PCRE in four places to do this. The result is
1773 dramatic:
1774
1775 Originally: 290
1776 After changing UCP table: 187
1777 After changing error message table: 43
1778 After changing table of "verbs" 36
1779 After changing table of Posix names 22
1780
1781 Thanks to the folks working on Gregex for glib for this insight.
1782
1783 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1784 unicode-properties was also set.
1785
1786 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1787
1788 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1789 checked only for CRLF.
1790
1791 14. Added casts to pcretest.c to avoid compiler warnings.
1792
1793 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1794
1795 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1796 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1797 entirely. This removes changes made in 7 above.
1798
1799 17. The CMake files have been updated, and there is now more information about
1800 building with CMake in the NON-UNIX-USE document.
1801
1802
1803 Version 7.3 28-Aug-07
1804 ---------------------
1805
1806 1. In the rejigging of the build system that eventually resulted in 7.1, the
1807 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1808 brackets there is not right, since it causes compilers to look for an
1809 installed pcre.h, not the version that is in the source that is being
1810 compiled (which of course may be different). I have changed it back to:
1811
1812 #include "pcre.h"
1813
1814 I have a vague recollection that the change was concerned with compiling in
1815 different directories, but in the new build system, that is taken care of
1816 by the VPATH setting the Makefile.
1817
1818 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1819 when the subject happened to end in the byte 0x85 (e.g. if the last
1820 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1821 characters but of course it shouldn't be taken as a newline when it is part
1822 of another character. The bug was that, for an unlimited repeat of . in
1823 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1824 characters when looking for a newline.
1825
1826 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1827
1828 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1829 in debug output.
1830
1831 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1832 long printing in the pcrecpp unittest when running under MinGW.
1833
1834 6. ESC_K was left out of the EBCDIC table.
1835
1836 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1837 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1838 limit also applies to "virtual nesting" when a pattern is recursive, and in
1839 this case 1000 isn't so big. I have been able to remove this limit at the
1840 expense of backing off one optimization in certain circumstances. Normally,
1841 when pcre_exec() would call its internal match() function recursively and
1842 immediately return the result unconditionally, it uses a "tail recursion"
1843 feature to save stack. However, when a subpattern that can match an empty
1844 string has an unlimited repetition quantifier, it no longer makes this
1845 optimization. That gives it a stack frame in which to save the data for
1846 checking that an empty string has been matched. Previously this was taken
1847 from the 1000-entry workspace that had been reserved. So now there is no
1848 explicit limit, but more stack is used.
1849
1850 8. Applied Daniel's patches to solve problems with the import/export magic
1851 syntax that is required for Windows, and which was going wrong for the
1852 pcreposix and pcrecpp parts of the library. These were overlooked when this
1853 problem was solved for the main library.
1854
1855 9. There were some crude static tests to avoid integer overflow when computing
1856 the size of patterns that contain repeated groups with explicit upper
1857 limits. As the maximum quantifier is 65535, the maximum group length was
1858 set at 30,000 so that the product of these two numbers did not overflow a
1859 32-bit integer. However, it turns out that people want to use groups that
1860 are longer than 30,000 bytes (though not repeat them that many times).
1861 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1862 made it possible to implement the integer overflow checks in a much more
1863 dynamic way, which I have now done. The artificial limitation on group
1864 length has been removed - we now have only the limit on the total length of
1865 the compiled pattern, which depends on the LINK_SIZE setting.
1866
1867 10. Fixed a bug in the documentation for get/copy named substring when
1868 duplicate names are permitted. If none of the named substrings are set, the
1869 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1870 empty string.
1871
1872 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1873 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1874 because the ] is interpreted as the first data character and the
1875 terminating ] is not found. PCRE has been made compatible with Perl in this
1876 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1877 cause memory overwriting.
1878
1879 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1880 string has been matched (to stop an infinite loop). It was not recognizing
1881 a conditional subpattern that could match an empty string if that
1882 subpattern was within another subpattern. For example, it looped when
1883 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1884 condition was not nested. This bug has been fixed.
1885
1886 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1887 past the start of the subject in the presence of bytes with the top bit
1888 set, for example "\x8aBCD".
1889
1890 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1891 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1892
1893 14. Optimized (?!) to (*FAIL).
1894
1895 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1896 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1897 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1898 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1899 does: it's just the validity check that is more restrictive.
1900
1901 16. Inserted checks for integer overflows during escape sequence (backslash)
1902 processing, and also fixed erroneous offset values for syntax errors during
1903 backslash processing.
1904
1905 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1906 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1907
1908 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1909 caused an overrun.
1910
1911 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1912 something other than just ASCII characters) inside a group that had an
1913 unlimited repeat caused a loop at compile time (while checking to see
1914 whether the group could match an empty string).
1915
1916 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1917 [\P{Any}] did so. (Error in the code for printing property names.)
1918
1919 21. An orphan \E inside a character class could cause a crash.
1920
1921 22. A repeated capturing bracket such as (A)? could cause a wild memory
1922 reference during compilation.
1923
1924 23. There are several functions in pcre_compile() that scan along a compiled
1925 expression for various reasons (e.g. to see if it's fixed length for look
1926 behind). There were bugs in these functions when a repeated \p or \P was
1927 present in the pattern. These operators have additional parameters compared
1928 with \d, etc, and these were not being taken into account when moving along
1929 the compiled data. Specifically:
1930
1931 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1932 length.
1933
1934 (b) An item such as \pL+ within a repeated group could cause crashes or
1935 loops.
1936
1937 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1938 "reference to non-existent subpattern" error.
1939
1940 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1941
1942 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1943 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1944
1945 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1946
1947 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1948 character were causing crashes (broken optimization).
1949
1950 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1951 \p or \P) caused a compile-time loop.
1952
1953 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1954 break. For example, the unstudied pattern [\r\n]A does not match the string
1955 "\r\nA" because change 7.0/46 below moves the current point on by two
1956 characters after failing to match at the start. However, the pattern \nA
1957 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1958 the same is true. There doesn't seem any very clean way out of this, but
1959 what I have chosen to do makes the common cases work: PCRE now takes note
1960 of whether there can be an explicit match for \r or \n anywhere in the
1961 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1962 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1963 pattern has explicit CR or LF references.
1964
1965 29. Added (*CR) etc for changing newline setting at start of pattern.
1966
1967
1968 Version 7.2 19-Jun-07
1969 ---------------------
1970
1971 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1972 which is apparently normally available under Windows.
1973
1974 2. Re-jig the pcregrep tests with different newline settings in an attempt
1975 to make them independent of the local environment's newline setting.
1976
1977 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1978
1979 4. Some of the "internals" tests were previously cut out when the link size
1980 was not 2, because the output contained actual offsets. The recent new
1981 "Z" feature of pcretest means that these can be cut out, making the tests
1982 usable with all link sizes.
1983
1984 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1985 stack recursion. This gives a massive performance boost under BSD, but just
1986 a small improvement under Linux. However, it saves one field in the frame
1987 in all cases.
1988
1989 6. Added more features from the forthcoming Perl 5.10:
1990
1991 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1992 recursion call. It refers to the nth most recently opened parentheses.
1993
1994 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1995 to be opened parentheses.
1996
1997 (c) Conditions that refer to capturing parentheses can be specified
1998 relatively, for example, (?(-2)... or (?(+3)...
1999
2000 (d) \K resets the start of the current match so that everything before
2001 is not part of it.
2002
2003 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2004
2005 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2006 reference syntax.
2007
2008 (g) (?| introduces a group in which the numbering of parentheses in each
2009 alternative starts with the same number.
2010
2011 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2012
2013 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2014 PCRE_INFO_JCHANGED.
2015
2016 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2017 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2018 for detecting groups that can match an empty string.
2019
2020 9. A pattern with a very large number of alternatives (more than several
2021 hundred) was running out of internal workspace during the pre-compile
2022 phase, where pcre_compile() figures out how much memory will be needed. A
2023 bit of new cunning has reduced the workspace needed for groups with
2024 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2025 workspace instead of running out of the 4096 that are available.
2026
2027 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2028
2029 11. Applied patch from Google to remove an optimization that didn't quite work.
2030 The report of the bug said:
2031
2032 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2033 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2034 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2035
2036 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2037 it matched the wrong number of bytes.
2038
2039
2040 Version 7.1 24-Apr-07
2041 ---------------------
2042
2043 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2044 that is more "standard", making use of automake and other Autotools. There
2045 is some re-arrangement of the files and adjustment of comments consequent
2046 on this.
2047
2048 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2049 for recursive directory scanning broke on some systems because the files
2050 are not scanned in any specific order and on different systems the order
2051 was different. A call to "sort" has been inserted into RunGrepTest for the
2052 approprate test as a short-term fix. In the longer term there may be an
2053 alternative.
2054
2055 3. I had an email from Eric Raymond about problems translating some of PCRE's
2056 man pages to HTML (despite the fact that I distribute HTML pages, some
2057 people do their own conversions for various reasons). The problems
2058 concerned the use of low-level troff macros .br and .in. I have therefore
2059 removed all such uses from the man pages (some were redundant, some could
2060 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2061 HTML has been updated to handle .nf/.fi and to complain if it encounters
2062 .br or .in.
2063
2064 4. Updated comments in configure.ac that get placed in config.h.in and also
2065 arranged for config.h to be included in the distribution, with the name
2066 config.h.generic, for the benefit of those who have to compile without
2067 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2068
2069 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2070 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2071 makevp.bat for the current PCRE, using the additional files
2072 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2073
2074 6. A Windows user reported a minor discrepancy with test 2, which turned out
2075 to be caused by a trailing space on an input line that had got lost in his
2076 copy. The trailing space was an accident, so I've just removed it.
2077
2078 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2079 that is needed.
2080
2081 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2082 as "const" (a) because they are and (b) because it helps the PHP
2083 maintainers who have recently made a script to detect big data structures
2084 in the php code that should be moved to the .rodata section. I remembered
2085 to update Builducptable as well, so it won't revert if ucptable.h is ever
2086 re-created.
2087
2088 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2089 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2090 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2091 support is not required. This saves 1.5-2K of code, which is important in
2092 some applications.
2093
2094 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2095 so as not to refer to the tables, even though these functions will never be
2096 called when UTF-8 support is disabled. Otherwise there are problems with a
2097 shared library.
2098
2099 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2100
2101 (a) It was defining its arguments as char * instead of void *.
2102
2103 (b) It was assuming that all moves were upwards in memory; this was true
2104 a long time ago when I wrote it, but is no longer the case.
2105
2106 The emulated memove() is provided for those environments that have neither
2107 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2108 is clearly not the case, as these two bugs were recently reported.
2109
2110 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2111 and Detrail to create the HTML documentation, the .txt form of the man
2112 pages, and it removes trailing spaces from listed files. It also creates
2113 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2114 case, it wraps all the #defines with #ifndefs. This script should be run
2115 before "make dist".
2116
2117 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2118 with Unicode property support.
2119
2120 (a) For a maximizing quantifier, if the two different cases of the
2121 character were of different lengths in their UTF-8 codings (there are
2122 some cases like this - I found 11), and the matching function had to
2123 back up over a mixture of the two cases, it incorrectly assumed they
2124 were both the same length.
2125
2126 (b) When PCRE was configured to use the heap rather than the stack for
2127 recursion during matching, it was not correctly preserving the data for
2128 the other case of a UTF-8 character when checking ahead for a match
2129 while processing a minimizing repeat. If the check also involved
2130 matching a wide character, but failed, corruption could cause an
2131 erroneous result when trying to check for a repeat of the original
2132 character.
2133
2134 13. Some tidying changes to the testing mechanism:
2135
2136 (a) The RunTest script now detects the internal link size and whether there
2137 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2138 values substituted by "configure". (The RunGrepTest script already did
2139 this for UTF-8.) The configure.ac script no longer substitutes the
2140 relevant variables.
2141
2142 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2143 with length and offset values. This means that the output is different
2144 for different internal link sizes. Test 2 is skipped for link sizes
2145 other than 2 because of this, bypassing the problem. Unfortunately,
2146 there was also a test in test 3 (the locale tests) that used /B and
2147 failed for link sizes other than 2. Rather than cut the whole test out,
2148 I have added a new /Z option to pcretest that replaces the length and
2149 offset values with spaces. This is now used to make test 3 independent
2150 of link size. (Test 2 will be tidied up later.)
2151
2152 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2153 segmentation fault instead of returning the appropriate error message.
2154
2155 15. In multiline mode when the newline sequence was set to "any", the pattern
2156 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2157 This doesn't seem right; it now treats the CRLF combination as the line
2158 ending, and so does not match in that case. It's only a pattern such as ^$
2159 that would hit this one: something like ^ABC$ would have failed after \r
2160 and then tried again after \r\n.
2161
2162 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2163 in an attempt to make files that differ only in their line terminators
2164 compare equal. This works on Linux.
2165
2166 17. Under certain error circumstances pcregrep might try to free random memory
2167 as it exited. This is now fixed, thanks to valgrind.
2168
2169 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2170 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2171 was because its rules for how to advance for /g after matching an empty
2172 string at the end of a line did not allow for this case. They now check for
2173 it specially.
2174
2175 20. pcretest is supposed to handle patterns and data of any length, by
2176 extending its buffers when necessary. It was getting this wrong when the
2177 buffer for a data line had to be extended.
2178
2179 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2180 CRLF as a newline sequence.
2181
2182 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2183 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2184 I have nevertheless tidied it up.
2185
2186 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2187
2188 24. Added a man page for pcre-config.
2189
2190
2191 Version 7.0 19-Dec-06
2192 ---------------------
2193
2194 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2195 moving to gcc 4.1.1.
2196
2197 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2198 sys/time.h, which is documented as needed for this function. It doesn't
2199 seem to matter on Linux, but it showed up on some releases of OS X.
2200
2201 3. It seems that there are systems where bytes whose values are greater than
2202 127 match isprint() in the "C" locale. The "C" locale should be the
2203 default when a C program starts up. In most systems, only ASCII printing
2204 characters match isprint(). This difference caused the output from pcretest
2205 to vary, making some of the tests fail. I have changed pcretest so that:
2206
2207 (a) When it is outputting text in the compiled version of a pattern, bytes
2208 other than 32-126 are always shown as hex escapes.
2209
2210 (b) When it is outputting text that is a matched part of a subject string,
2211 it does the same, unless a different locale has been set for the match
2212 (using the /L modifier). In this case, it uses isprint() to decide.
2213
2214 4. Fixed a major bug that caused incorrect computation of the amount of memory
2215 required for a compiled pattern when options that changed within the
2216 pattern affected the logic of the preliminary scan that determines the
2217 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2218 that the computed length was too small. The symptoms of this bug were
2219 either the PCRE error "internal error: code overflow" from pcre_compile(),
2220 or a glibc crash with a message such as "pcretest: free(): invalid next
2221 size (fast)". Examples of patterns that provoked this bug (shown in
2222 pcretest format) are:
2223
2224 /(?-x: )/x
2225 /(?x)(?-x: \s*#\s*)/
2226 /((?i)[\x{c0}])/8
2227 /(?i:[\x{c0}])/8
2228
2229 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2230 is now done differently.
2231
2232 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2233 wrapper classes; (b) implement a new function in the C++ scanner that is
2234 more efficient than the old way of doing things because it avoids levels of
2235 recursion in the regex matching; (c) add a paragraph to the documentation
2236 for the FullMatch() function.
2237
2238 6. The escape sequence \n was being treated as whatever was defined as
2239 "newline". Not only was this contrary to the documentation, which states
2240 that \n is character 10 (hex 0A), but it also went horribly wrong when
2241 "newline" was defined as CRLF. This has been fixed.
2242
2243 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2244 was being set to -1 for the "end of line" case (supposedly a value that no
2245 character can have). Though this value is never used (the check for end of
2246 line is "zero bytes in current character"), it caused compiler complaints.
2247 I've changed it to 0xffffffff.
2248
2249 8. In pcre_version.c, the version string was being built by a sequence of
2250 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2251 string (as it is for production releases) called a macro with an empty
2252 argument. The C standard says the result of this is undefined. The gcc
2253 compiler treats it as an empty string (which was what was wanted) but it is
2254 reported that Visual C gives an error. The source has been hacked around to
2255 avoid this problem.
2256
2257 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2258 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2259 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2260 of them did).
2261
2262 10. Originally, pcretest opened its input and output without "b"; then I was
2263 told that "b" was needed in some environments, so it was added for release
2264 5.0 to both the input and output. (It makes no difference on Unix-like
2265 systems.) Later I was told that it is wrong for the input on Windows. I've
2266 now abstracted the modes into two macros, to make it easier to fiddle with
2267 them, and removed "b" from the input mode under Windows.
2268
2269 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2270
2271 12. Added -help and --help to pcretest as an official way of being reminded
2272 of the options.
2273
2274 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2275 and pcrecpp.cc because they annoy compilers at high warning levels.
2276
2277 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2278
2279 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2280 scripts are not C programs :-) and which was not noticed because it works
2281 on Linux.
2282
2283 16. pcretest is supposed to handle any length of pattern and data line (as one
2284 line or as a continued sequence of lines) by extending its input buffer if
2285 necessary. This feature was broken for very long pattern lines, leading to
2286 a string of junk being passed to pcre_compile() if the pattern was longer
2287 than about 50K.
2288
2289 17. I have done a major re-factoring of the way pcre_compile() computes the
2290 amount of memory needed for a compiled pattern. Previously, there was code
2291 that made a preliminary scan of the pattern in order to do this. That was
2292 OK when PCRE was new, but as the facilities have expanded, it has become
2293 harder and harder to keep it in step with the real compile phase, and there
2294 have been a number of bugs (see for example, 4 above). I have now found a
2295 cunning way of running the real compile function in a "fake" mode that
2296 enables it to compute how much memory it would need, while actually only
2297 ever using a few hundred bytes of working memory and without too many
2298 tests of the mode. This should make future maintenance and development
2299 easier. A side effect of this work is that the limit of 200 on the nesting
2300 depth of parentheses has been removed (though this was never a serious
2301 limitation, I suspect). However, there is a downside: pcre_compile() now
2302 runs more slowly than before (30% or more, depending on the pattern). I
2303 hope this isn't a big issue. There is no effect on runtime performance.
2304
2305 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2306 newline (only possible for the last line of a file) and it was a
2307 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2308
2309 19. Added additional timing features to pcretest. (1) The -tm option now times
2310 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2311 separate command line item, by a number that specifies the number of
2312 repeats to use when timing. The default is 50000; this gives better
2313 precision, but takes uncomfortably long for very large patterns.
2314
2315 20. Extended pcre_study() to be more clever in cases where a branch of a
2316 subpattern has no definite first character. For example, (a*|b*)[cd] would
2317 previously give no result from pcre_study(). Now it recognizes that the
2318 first character must be a, b, c, or d.
2319
2320 21. There was an incorrect error "recursive call could loop indefinitely" if
2321 a subpattern (or the entire pattern) that was being tested for matching an
2322 empty string contained only one non-empty item after a nested subpattern.
2323 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2324 incorrectly, because the \d was being skipped in the check.
2325
2326 22. The pcretest program now has a new pattern option /B and a command line
2327 option -b, which is equivalent to adding /B to every pattern. This causes
2328 it to show the compiled bytecode, without the additional information that
2329 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2330 is the same as /B/I).
2331
2332 23. A new optimization is now able automatically to treat some sequences such
2333 as a*b as a*+b. More specifically, if something simple (such as a character
2334 or a simple class like \d) has an unlimited quantifier, and is followed by
2335 something that cannot possibly match the quantified thing, the quantifier
2336 is automatically "possessified".
2337
2338 24. A recursive reference to a subpattern whose number was greater than 39
2339 went wrong under certain circumstances in UTF-8 mode. This bug could also
2340 have affected the operation of pcre_study().
2341
2342 25. Realized that a little bit of performance could be had by replacing
2343 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2344
2345 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2346
2347 27. Possessive quantifiers such as a++ were previously implemented by turning
2348 them into atomic groups such as ($>a+). Now they have their own opcodes,
2349 which improves performance. This includes the automatically created ones
2350 from 23 above.
2351
2352 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2353 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2354 the first matched character to be a colon. This applied both to named and
2355 numbered groups.
2356
2357 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2358
2359 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2360 building PCRE on Windows easier, so I have included it in the distribution.
2361
2362 31. There is now a check in pcretest against a ridiculously large number being
2363 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2364 loop, the loop is abandoned.
2365
2366 32. Forward references to subpatterns in conditions such as (?(2)...) where
2367 subpattern 2 is defined later cause pcre_compile() to search forwards in
2368 the pattern for the relevant set of parentheses. This search went wrong
2369 when there were unescaped parentheses in a character class, parentheses
2370 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2371
2372 33. "Subroutine" calls and backreferences were previously restricted to
2373 referencing subpatterns earlier in the regex. This restriction has now
2374 been removed.
2375
2376 34. Added a number of extra features that are going to be in Perl 5.10. On the
2377 whole, these are just syntactic alternatives for features that PCRE had
2378 previously implemented using the Python syntax or my own invention. The
2379 other formats are all retained for compatibility.
2380
2381 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2382 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2383 also .NET compatible.
2384
2385 (b) A recursion or subroutine call to a named group can now be defined as
2386 (?&name) as well as (?P>name).
2387
2388 (c) A backreference to a named group can now be defined as \k<name> or
2389 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2390 5.10, are also .NET compatible.
2391
2392 (d) A conditional reference to a named group can now use the syntax
2393 (?(<name>) or (?('name') as well as (?(name).
2394
2395 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2396 groups (named and numbered) that are never evaluated inline, but can be
2397 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2398 is always false. There may be only one alternative in such a group.
2399
2400 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2401 as the simple (?(R). The condition is true only if the most recent
2402 recursion is that of the given number or name. It does not search out
2403 through the entire recursion stack.
2404
2405 (g) The escape \gN or \g{N} has been added, where N is a positive or
2406 negative number, specifying an absolute or relative reference.
2407
2408 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2409 some "unreachable code" warnings.
2410
2411 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2412 things, this adds five new scripts.
2413
2414 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2415 There were also incompatibilities regarding the handling of \Q..\E inside
2416 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2417 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2418
2419 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2420 matches an empty string, and forcibly breaks the loop. There were bugs in
2421 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2422 against aaaa the result was just "a" rather than "aaaa", for example. Two
2423 separate and independent bugs (that affected different cases) have been
2424 fixed.
2425
2426 39. Refactored the code to abolish the use of different opcodes for small
2427 capturing bracket numbers. This is a tidy that I avoided doing when I
2428 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2429 The new approach is not only tidier, it makes it possible to reduce the
2430 memory needed to fix the previous bug (38).
2431
2432 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2433 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2434 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2435 mode.
2436
2437 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2438 report.
2439
2440 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2441 copy construction and assignment in the C++ wrapper.
2442
2443 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2444 couple of bugs that could have given wrong results in the "--newline=crlf"
2445 case.
2446
2447 44. Added a number of casts and did some reorganization of signed/unsigned int
2448 variables following suggestions from Dair Grant. Also renamed the variable
2449 "this" as "item" because it is a C++ keyword.
2450
2451 45. Arranged for dftables to add
2452
2453 #include "pcre_internal.h"
2454
2455 to pcre_chartables.c because without it, gcc 4.x may remove the array
2456 definition from the final binary if PCRE is built into a static library and
2457 dead code stripping is activated.
2458
2459 46. For an unanchored pattern, if a match attempt fails at the start of a
2460 newline sequence, and the newline setting is CRLF or ANY, and the next two
2461 characters are CRLF, advance by two characters instead of one.
2462
2463
2464 Version 6.7 04-Jul-06
2465 ---------------------
2466
2467 1. In order to handle tests when input lines are enormously long, pcretest has
2468 been re-factored so that it automatically extends its buffers when
2469 necessary. The code is crude, but this _is_ just a test program. The
2470 default size has been increased from 32K to 50K.
2471
2472 2. The code in pcre_study() was using the value of the re argument before
2473 testing it for NULL. (Of course, in any sensible call of the function, it
2474 won't be NULL.)
2475
2476 3. The memmove() emulation function in pcre_internal.h, which is used on
2477 systems that lack both memmove() and bcopy() - that is, hardly ever -
2478 was missing a "static" storage class specifier.
2479
2480 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2481 containing an extended class (one that cannot be represented by a bitmap
2482 because it contains high-valued characters or Unicode property items, e.g.
2483 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2484 pattern, but PCRE should not loop if you do not (it no longer does).
2485 [Detail: two cases were found: (a) a repeated subpattern containing an
2486 extended class; (b) a recursive reference to a subpattern that followed a
2487 previous extended class. It wasn't skipping over the extended class
2488 correctly when UTF-8 mode was not set.]
2489
2490 5. A negated single-character class was not being recognized as fixed-length
2491 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2492 compile error "lookbehind assertion is not fixed length".
2493
2494 6. The RunPerlTest auxiliary script was showing an unexpected difference
2495 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2496 write a Perl script that can interpret lines of an input file either as
2497 byte characters or as UTF-8, which is what "perltest" was being required to
2498 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2499 can't do is switch easily at run time between having the "use utf8;" pragma
2500 or not. In the end, I fudged it by using the RunPerlTest script to insert
2501 "use utf8;" explicitly for the UTF-8 tests.
2502
2503 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2504 the end of the subject string, contrary to the documentation and to what
2505 Perl does. This was true of both matching functions. Now it matches only at
2506 the start of the subject and immediately after *internal* newlines.
2507
2508 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2509 a pointer to an int instead of a pointer to an unsigned long int. This
2510 caused problems on 64-bit systems.
2511
2512 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2513 instance of the 'standard' template library not being so standard".
2514
2515 10. There was no check on the number of named subpatterns nor the maximum
2516 length of a subpattern name. The product of these values is used to compute
2517 the size of the memory block for a compiled pattern. By supplying a very
2518 long subpattern name and a large number of named subpatterns, the size
2519 computation could be caused to overflow. This is now prevented by limiting
2520 the length of names to 32 characters, and the number of named subpatterns
2521 to 10,000.
2522
2523 11. Subpatterns that are repeated with specific counts have to be replicated in
2524 the compiled pattern. The size of memory for this was computed from the
2525 length of the subpattern and the repeat count. The latter is limited to
2526 65535, but there was no limit on the former, meaning that integer overflow
2527 could in principle occur. The compiled length of a repeated subpattern is
2528 now limited to 30,000 bytes in order to prevent this.
2529
2530 12. Added the optional facility to have named substrings with the same name.
2531
2532 13. Added the ability to use a named substring as a condition, using the
2533 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2534 are numbers (not recommended). Forward references are permitted.
2535
2536 14. Added forward references in named backreferences (if you see what I mean).
2537
2538 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2539 pattern could run off the end of the subject. For example, the pattern
2540 "(?s)(.{1,5})"8 did this with the subject "ab".
2541
2542 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2543 PCRE_CASELESS was set when matching characters that were quantified with ?
2544 or *.
2545
2546 17. A character class other than a single negated character that had a minimum
2547 but no maximum quantifier - for example [ab]{6,} - was not handled
2548 correctly by pce_dfa_exec(). It would match only one character.
2549
2550 18. A valid (though odd) pattern that looked like a POSIX character
2551 class but used an invalid character after [ (for example [[,abc,]]) caused
2552 pcre_compile() to give the error "Failed: internal error: code overflow" or
2553 in some cases to crash with a glibc free() error. This could even happen if
2554 the pattern terminated after [[ but there just happened to be a sequence of
2555 letters, a binary zero, and a closing ] in the memory that followed.
2556
2557 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2558 over the years. Originally (before any Unicode support), just the bottom 8
2559 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2560 output from "man perlunicode" includes this:
2561
2562 The regular expression compiler produces polymorphic opcodes. That
2563 is, the pattern adapts to the data and automatically switches to
2564 the Unicode character scheme when presented with Unicode data--or
2565 instead uses a traditional byte scheme when presented with byte
2566 data.
2567
2568 Sadly, a wide octal escape does not cause a switch, and in a string with
2569 no other multibyte characters, these octal escapes are treated as before.
2570 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2571 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2572 Unicode string.
2573
2574 I have not perpetrated such confusion in PCRE. Up till now, it took just
2575 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2576 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2577 translate to the appropriate multibyte character.
2578
2579 29. Applied some refactoring to reduce the number of warnings from Microsoft
2580 and Borland compilers. This has included removing the fudge introduced
2581 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2582 a warning about an unused variable.
2583
2584 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2585 characters since release 4.0, because Perl (from release 5.004) does not.
2586 [Or at least, is documented not to: some releases seem to be in conflict
2587 with the documentation.] However, when a pattern was studied with
2588 pcre_study() and all its branches started with \s, PCRE still included VT
2589 as a possible starting character. Of course, this did no harm; it just
2590 caused an unnecessary match attempt.
2591
2592 22. Removed a now-redundant internal flag bit that recorded the fact that case
2593 dependency changed within the pattern. This was once needed for "required
2594 byte" processing, but is no longer used. This recovers a now-scarce options
2595 bit. Also moved the least significant internal flag bit to the most-
2596 significant bit of the word, which was not previously used (hangover from
2597 the days when it was an int rather than a uint) to free up another bit for
2598 the future.
2599
2600 23. Added support for CRLF line endings as well as CR and LF. As well as the
2601 default being selectable at build time, it can now be changed at runtime
2602 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2603 specify that it is scanning data with non-default line endings.
2604
2605 24. Changed the definition of CXXLINK to make it agree with the definition of
2606 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2607
2608 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2609 recursions. This makes a big different to stack usage for some patterns.
2610
2611 26. If a subpattern containing a named recursion or subroutine reference such
2612 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2613 the space required for the compiled pattern went wrong and gave too small a
2614 value. Depending on the environment, this could lead to "Failed: internal
2615 error: code overflow at offset 49" or "glibc detected double free or
2616 corruption" errors.
2617
2618 27. Applied patches from Google (a) to support the new newline modes and (b) to
2619 advance over multibyte UTF-8 characters in GlobalReplace.
2620
2621 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2622 difference for some implementation of PCRE in some Windows version.
2623
2624 29. Added some extra testing facilities to pcretest:
2625
2626 \q<number> in a data line sets the "match limit" value
2627 \Q<number> in a data line sets the "match recursion limt" value
2628 -S <number> sets the stack size, where <number> is in megabytes
2629
2630 The -S option isn't available for Windows.
2631
2632
2633 Version 6.6 06-Feb-06
2634 ---------------------
2635
2636 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2637 in pcreposix.h. I have copied the definition from pcre.h.
2638
2639 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2640 because pcre.h is no longer a built file.
2641
2642 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2643 not normally included in the compiled code.
2644
2645
2646 Version 6.5 01-Feb-06
2647 ---------------------
2648
2649 1. When using the partial match feature with pcre_dfa_exec(), it was not
2650 anchoring the second and subsequent partial matches at the new starting
2651 point. This could lead to incorrect results. For example, with the pattern
2652 /1234/, partially matching against "123" and then "a4" gave a match.
2653
2654 2. Changes to pcregrep:
2655
2656 (a) All non-match returns from pcre_exec() were being treated as failures
2657 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2658 error message is output. Some extra information is given for the
2659 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2660 probably the only errors that are likely to be caused by users (by
2661 specifying a regex that has nested indefinite repeats, for instance).
2662 If there are more than 20 of these errors, pcregrep is abandoned.
2663
2664 (b) A binary zero was treated as data while matching, but terminated the
2665 output line if it was written out. This has been fixed: binary zeroes
2666 are now no different to any other data bytes.
2667
2668 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2669 used to set a locale for matching. The --locale=xxxx long option has
2670 been added (no short equivalent) to specify a locale explicitly on the
2671 pcregrep command, overriding the environment variables.
2672
2673 (d) When -B was used with -n, some line numbers in the output were one less
2674 than they should have been.
2675
2676 (e) Added the -o (--only-matching) option.
2677
2678 (f) If -A or -C was used with -c (count only), some lines of context were
2679 accidentally printed for the final match.
2680
2681 (g) Added the -H (--with-filename) option.
2682
2683 (h) The combination of options -rh failed to suppress file names for files
2684 that were found from directory arguments.
2685
2686 (i) Added the -D (--devices) and -d (--directories) options.
2687
2688 (j) Added the -F (--fixed-strings) option.
2689
2690 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2691
2692 (l) Added the --colo(u)r option.
2693
2694 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2695 is not present by default.
2696
2697 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2698 items such as (?R) or (?1), when the recursion could match a number of
2699 alternatives. If it matched one of the alternatives, but subsequently,
2700 outside the recursion, there was a failure, the code tried to back up into
2701 the recursion. However, because of the way PCRE is implemented, this is not
2702 possible, and the result was an incorrect result from the match.
2703
2704 In order to prevent this happening, the specification of recursion has
2705 been changed so that all such subpatterns are automatically treated as
2706 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2707
2708 4. I had overlooked the fact that, in some locales, there are characters for
2709 which isalpha() is true but neither isupper() nor islower() are true. In
2710 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2711 and ordfeminine) are like this. This affected the treatment of \w and \W
2712 when they appeared in character classes, but not when they appeared outside
2713 a character class. The bit map for "word" characters is now created
2714 separately from the results of isalnum() instead of just taking it from the
2715 upper, lower, and digit maps. (Plus the underscore character, of course.)
2716
2717 5. The above bug also affected the handling of POSIX character classes such as
2718 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2719 permanent tables. Instead, the bit maps for such a class were previously
2720 created as the appropriate unions of the upper, lower, and digit bitmaps.
2721 Now they are created by subtraction from the [[:word:]] class, which has
2722 its own bitmap.
2723
2724 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2725 It is created by subtracting the vertical space characters (\x09, \x0a,
2726 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2727 subtraction was done in the overall bitmap for a character class, meaning
2728 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2729 be recognized. This bug has been fixed.
2730
2731 7. Patches from the folks at Google:
2732
2733 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2734 real life, but is still worth protecting against".
2735
2736 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2737 regular expressions".
2738
2739 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2740 have it.
2741
2742 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2743 "configure" and the latter not, in order to fix a problem somebody had
2744 with compiling the Arg class on HP-UX.
2745
2746 (e) Improve the error-handling of the C++ wrapper a little bit.
2747
2748 (f) New tests for checking recursion limiting.
2749
2750 8. The pcre_memmove() function, which is used only if the environment does not
2751 have a standard memmove() function (and is therefore rarely compiled),
2752 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2753 returning a result (though PCRE never actually uses the result).
2754
2755 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2756 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2757 returned instead of calling malloc() with an overflowing number that would
2758 most likely cause subsequent chaos.
2759
2760 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2761
2762 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2763 with this option is matched, the nmatch and pmatch options of regexec() are
2764 ignored.
2765
2766 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2767 provided in case anyone wants to the the POSIX interface with UTF-8
2768 strings.
2769
2770 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2771 C++ linking (needed for some HP-UX environments).
2772
2773 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2774 (unused parameter) and in the pcre_printint() function (omitted "default"
2775 switch label when the default is to do nothing).
2776
2777 15. Added some code to make it possible, when PCRE is compiled as a C++
2778 library, to replace subject pointers for pcre_exec() with a smart pointer
2779 class, thus making it possible to process discontinuous strings.
2780
2781 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2782 much the same function. They were added by different people who were trying
2783 to make PCRE easy to compile on non-Unix systems. It has been suggested
2784 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2785 for compiling on Windows systems. I have therefore replaced it with
2786 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2787 defaults to "extern" for C or "extern C" for C++, which works fine on
2788 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2789 SCOPE with something explicit in config.h. In addition:
2790
2791 (a) pcreposix.h still had just "extern" instead of either of these macros;
2792 I have replaced it with PCRE_DATA_SCOPE.
2793
2794 (b) Functions such as _pcre_xclass(), which are internal to the library,
2795 but external in the C sense, all had PCRE_EXPORT in their definitions.
2796 This is apparently wrong for the Windows case, so I have removed it.
2797 (It makes no difference on Unix-like systems.)
2798
2799 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2800 of recursive calls to match(). This is different to MATCH_LIMIT because
2801 that limits the total number of calls to match(), not all of which increase
2802 the depth of recursion. Limiting the recursion depth limits the amount of
2803 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2804 when PCRE is compiled, and changed at run time. A patch from Google adds
2805 this functionality to the C++ interface.
2806
2807 18. Changes to the handling of Unicode character properties:
2808
2809 (a) Updated the table to Unicode 4.1.0.
2810
2811 (b) Recognize characters that are not in the table as "Cn" (undefined).
2812
2813 (c) I revised the way the table is implemented to a much improved format
2814 which includes recognition of ranges. It now supports the ranges that
2815 are defined in UnicodeData.txt, and it also amalgamates other
2816 characters into ranges. This has reduced the number of entries in the
2817 table from around 16,000 to around 3,000, thus reducing its size
2818 considerably. I realized I did not need to use a tree structure after
2819 all - a binary chop search is just as efficient. Having reduced the
2820 number of entries, I extended their size from 6 bytes to 8 bytes to
2821 allow for more data.
2822
2823 (d) Added support for Unicode script names via properties such as \p{Han}.
2824
2825 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2826 matching that character.
2827
2828 20. When matching a repeated Unicode property with a minimum greater than zero,
2829 (for example \pL{2,}), PCRE could look past the end of the subject if it
2830 reached it while seeking the minimum number of characters. This could
2831 happen only if some of the characters were more than one byte long, because
2832 there is a check for at least the minimum number of bytes.
2833
2834 21. Refactored the implementation of \p and \P so as to be more general, to
2835 allow for more different types of property in future. This has changed the
2836 compiled form incompatibly. Anybody with saved compiled patterns that use
2837 \p or \P will have to recompile them.
2838
2839 22. Added "Any" and "L&" to the supported property types.
2840
2841 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2842 but give a compile time error if the value is greater than 0xff.
2843
2844 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2845 accidentally not being installed or uninstalled.
2846
2847 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2848 made were to insert the current release number. This seemed silly, because
2849 it made things harder for people building PCRE on systems that don't run
2850 "configure". I have turned pcre.h into a distributed file, no longer built
2851 by "configure", with the version identification directly included. There is
2852 no longer a pcre.h.in file.
2853
2854 However, this change necessitated a change to the pcre-config script as
2855 well. It is built from pcre-config.in, and one of the substitutions was the
2856 release number. I have updated configure.ac so that ./configure now finds
2857 the release number by grepping pcre.h.
2858
2859 26. Added the ability to run the tests under valgrind.
2860
2861
2862 Version 6.4 05-Sep-05
2863 ---------------------
2864
2865 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2866 "--" to be printed when multiple files were scanned, even when none of the
2867 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2868 consider it to be a bug, and have restored the previous behaviour.
2869
2870 2. A couple of code tidies to get rid of compiler warnings.
2871
2872 3. The pcretest program used to cheat by referring to symbols in the library
2873 whose names begin with _pcre_. These are internal symbols that are not
2874 really supposed to be visible externally, and in some environments it is
2875 possible to suppress them. The cheating is now confined to including
2876 certain files from the library's source, which is a bit cleaner.
2877
2878 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2879 file's purpose clearer.
2880
2881 5. Reorganized pcre_ucp_findchar().
2882
2883
2884 Version 6.3 15-Aug-05
2885 ---------------------
2886
2887 1. The file libpcre.pc.in did not have general read permission in the tarball.
2888
2889 2. There were some problems when building without C++ support:
2890
2891 (a) If C++ support was not built, "make install" and "make test" still
2892 tried to test it.
2893
2894 (b) There were problems when the value of CXX was explicitly set. Some
2895 changes have been made to try to fix these, and ...
2896
2897 (c) --disable-cpp can now be used to explicitly disable C++ support.
2898
2899 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2900 backslash in a target when C++ was disabled. This confuses some
2901 versions of "make", apparently. Using an intermediate variable solves
2902 this. (Same for CPP_LOBJ.)
2903
2904 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2905 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2906 necessary on certain architectures.
2907
2908 4. Added a setting of -export-symbols-regex to the link command to remove
2909 those symbols that are exported in the C sense, but actually are local
2910 within the library, and not documented. Their names all begin with
2911 "_pcre_". This is not a perfect job, because (a) we have to except some
2912 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2913 available (and never for static libraries). I have made a note to try to
2914 find a way round (a) in the future.
2915
2916
2917 Version 6.2 01-Aug-05
2918 ---------------------
2919
2920 1. There was no test for integer overflow of quantifier values. A construction
2921 such as {1111111111111111} would give undefined results. What is worse, if
2922 a minimum quantifier for a parenthesized subpattern overflowed and became
2923 negative, the calculation of the memory size went wrong. This could have
2924 led to memory overwriting.
2925
2926 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2927
2928 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2929 operating environments where this matters.
2930
2931 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2932 PCRE options from within the C++ wrapper.
2933
2934 5. Named capturing subpatterns were not being correctly counted when a pattern
2935 was compiled. This caused two problems: (a) If there were more than 100
2936 such subpatterns, the calculation of the memory needed for the whole
2937 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2938 back references of the form \12, where the number was greater than 9, were
2939 not recognized as back references, even though there were sufficient
2940 previous subpatterns.
2941
2942 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2943 versions of gcc, e.g. 2.95.4.
2944
2945
2946 Version 6.1 21-Jun-05
2947 ---------------------
2948
2949 1. There was one reference to the variable "posix" in pcretest.c that was not
2950 surrounded by "#if !defined NOPOSIX".
2951
2952 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2953 the cross-check on the old pcre_info() function, for the benefit of the
2954 cut-down version of PCRE that is currently imported into Exim.
2955
2956 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2957 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2958 patterns that start (?i) but I don't think that matters. The use of (?i) is
2959 just an example; this all applies to the other options as well.
2960
2961 4. Since libtool seems to echo the compile commands it is issuing, the output
2962 from "make" can be reduced a bit by putting "@" in front of each libtool
2963 compile command.
2964
2965 5. Patch from the folks at Google for configure.in to be a bit more thorough
2966 in checking for a suitable C++ installation before trying to compile the
2967 C++ stuff. This should fix a reported problem when a compiler was present,
2968 but no suitable headers.
2969
2970 6. The man pages all had just "PCRE" as their title. I have changed them to
2971 be the relevant file name. I have also arranged that these names are
2972 retained in the file doc/pcre.txt, which is a concatenation in text format
2973 of all the man pages except the little individual ones for each function.
2974
2975 7. The NON-UNIX-USE file had not been updated for the different set of source
2976 files that come with release 6. I also added a few comments about the C++
2977 wrapper.
2978
2979
2980 Version 6.0 07-Jun-05
2981 ---------------------
2982
2983 1. Some minor internal re-organization to help with my DFA experiments.
2984
2985 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2986 didn't matter for the library itself when fully configured, but did matter
2987 when compiling without UCP support, or within Exim, where the ucp files are
2988 not imported.
2989
2990 3. Refactoring of the library code to split up the various functions into
2991 different source modules. The addition of the new DFA matching code (see
2992 below) to a single monolithic source would have made it really too
2993 unwieldy, quite apart from causing all the code to be include in a
2994 statically linked application, when only some functions are used. This is
2995 relevant even without the DFA addition now that patterns can be compiled in
2996 one application and matched in another.
2997
2998 The downside of splitting up is that there have to be some external
2999 functions and data tables that are used internally in different modules of
3000 the library but which are not part of the API. These have all had their
3001 names changed to start with "_pcre_" so that they are unlikely to clash
3002 with other external names.
3003
3004 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3005 a different (DFA) algorithm. Although it is slower than the original
3006 function, it does have some advantages for certain types of matching
3007 problem.
3008
3009 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3010 including restarting after a partial match.
3011
3012 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3013 defined when compiling for Windows was sent to me. I have put it into the
3014 code, though I have no means of testing or verifying it.
3015
3016 7. Added the pcre_refcount() auxiliary function.
3017
3018 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3019 match before or at the first newline in the subject string. In pcretest,
3020 the /f option on a pattern can be used to set this.
3021
3022 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3023 would behave wrongly. This has been present in PCRE since release 4.0.
3024
3025 10. A number of changes to the pcregrep command:
3026
3027 (a) Refactored how -x works; insert ^(...)$ instead of setting
3028 PCRE_ANCHORED and checking the length, in preparation for adding
3029 something similar for -w.
3030
3031 (b) Added the -w (match as a word) option.
3032
3033 (c) Refactored the way lines are read and buffered so as to have more
3034 than one at a time available.
3035
3036 (d) Implemented a pcregrep test script.
3037
3038 (e) Added the -M (multiline match) option. This allows patterns to match
3039 over several lines of the subject. The buffering ensures that at least
3040 8K, or the rest of the document (whichever is the shorter) is available
3041 for matching (and similarly the previous 8K for lookbehind assertions).
3042
3043 (f) Changed the --help output so that it now says
3044
3045 -w, --word-regex(p)
3046
3047 instead of two lines, one with "regex" and the other with "regexp"
3048 because that confused at least one person since the short forms are the
3049 same. (This required a bit of code, as the output is generated
3050 automatically from a table. It wasn't just a text change.)
3051
3052 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3053 option but starts with a hyphen. Could be a pattern or a path name
3054 starting with a hyphen, for instance.
3055
3056 (h) "-" can be given as a file name to represent stdin.
3057
3058 (i) When file names are being printed, "(standard input)" is used for
3059 the standard input, for compatibility with GNU grep. Previously
3060 "<stdin>" was used.
3061
3062 (j) The option --label=xxx can be used to supply a name to be used for
3063 stdin when file names are being printed. There is no short form.
3064
3065 (k) Re-factored the options decoding logic because we are going to add
3066 two more options that take data. Such options can now be given in four
3067 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3068
3069 (l) Added the -A, -B, and -C options for requesting that lines of context
3070 around matches be printed.
3071
3072 (m) Added the -L option to print the names of files that do not contain
3073 any matching lines, that is, the complement of -l.
3074
3075 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3076 continue to scan other files.
3077
3078 (o) The -s option was incorrectly implemented. For compatibility with other
3079 greps, it now suppresses the error message for a non-existent or non-
3080 accessible file (but not the return code). There is a new option called
3081 -q that suppresses the output of matching lines, which was what -s was
3082 previously doing.
3083
3084 (p) Added --include and --exclude options to specify files for inclusion
3085 and exclusion when recursing.
3086
3087 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3088 Hopefully, it now does.
3089
3090 12. Missing cast in pcre_study().
3091
3092 13. Added an "uninstall" target to the makefile.
3093
3094 14. Replaced "extern" in the function prototypes in Makefile.in with
3095 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3096 world, but is set differently for Windows.
3097
3098 15. Added a second compiling function called pcre_compile2(). The only
3099 difference is that it has an extra argument, which is a pointer to an
3100 integer error code. When there is a compile-time failure, this is set
3101 non-zero, in addition to the error test pointer being set to point to an
3102 error message. The new argument may be NULL if no error number is required
3103 (but then you may as well call pcre_compile(), which is now just a
3104 wrapper). This facility is provided because some applications need a
3105 numeric error indication, but it has also enabled me to tidy up the way
3106 compile-time errors are handled in the POSIX wrapper.
3107
3108 16. Added VPATH=.libs to the makefile; this should help when building with one
3109 prefix path and installing with another. (Or so I'm told by someone who
3110 knows more about this stuff than I do.)
3111
3112 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3113 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3114 match everything, including newlines. This is not POSIX-compatible, but
3115 somebody wanted the feature. From pcretest it can be activated by using
3116 both the P and the s flags.
3117
3118 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3119
3120 19. libpcre.pc was being incorrectly installed as executable.
3121
3122 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3123 it now also looks for '\r' so that it will work unmodified on Windows.
3124
3125 21. Added Google's contributed C++ wrapper to the distribution.
3126
3127 22. Added some untidy missing memory free() calls in pcretest, to keep
3128 Electric Fence happy when testing.
3129
3130
3131
3132 Version 5.0 13-Sep-04
3133 ---------------------
3134
3135 1. Internal change: literal characters are no longer packed up into items
3136 containing multiple characters in a single byte-string. Each character
3137 is now matched using a separate opcode. However, there may be more than one
3138 byte in the character in UTF-8 mode.
3139
3140 2. The pcre_callout_block structure has two new fields: pattern_position and
3141 next_item_length. These contain the offset in the pattern to the next match
3142 item, and its length, respectively.
3143
3144 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3145 insertion of callouts before each pattern item. Added the /C option to
3146 pcretest to make use of this.
3147
3148 4. On the advice of a Windows user, the lines
3149
3150 #if defined(_WIN32) || defined(WIN32)
3151 _setmode( _fileno( stdout ), 0x8000 );
3152 #endif /* defined(_WIN32) || defined(WIN32) */
3153
3154 have been added to the source of pcretest. This apparently does useful
3155 magic in relation to line terminators.
3156
3157 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3158 for the benefit of those environments where the "b" makes a difference.
3159
3160 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3161 to know about it. I have put a hack into configure.in that adds in code
3162 to set GCC=yes if CC=icc. This seems to end up at a point in the
3163 generated configure script that is early enough to affect the setting of
3164 compiler options, which is what is needed, but I have no means of testing
3165 whether it really works. (The user who reported this had patched the
3166 generated configure script, which of course I cannot do.)
3167
3168 LATER: After change 22 below (new libtool files), the configure script
3169 seems to know about icc (and also ecc). Therefore, I have commented out
3170 this hack in configure.in.
3171
3172 7. Added support for pkg-config (2 patches were sent in).
3173
3174 8. Negated POSIX character classes that used a combination of internal tables
3175 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3176 [[:^ascii]]. Typically, they would match almost any characters. The other
3177 POSIX classes were not broken in this way.
3178
3179 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3180 to find the match, as PCRE was deluded into thinking that the match had to
3181 start at the start point or following a newline. The same bug applied to
3182 patterns with negative forward assertions or any backward assertions
3183 preceding ".*" at the start, unless the pattern required a fixed first
3184 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3185
3186 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3187 starting at the last subject character, bytes beyond the end of the subject
3188 string were read.
3189
3190 11. Renamed the variable "class" as "classbits" to make life easier for C++
3191 users. (Previously there was a macro definition, but it apparently wasn't
3192 enough.)
3193
3194 12. Added the new field "tables" to the extra data so that tables can be passed
3195 in at exec time, or the internal tables can be re-selected. This allows
3196 a compiled regex to be saved and re-used at a later time by a different
3197 program that might have everything at different addresses.
3198
3199 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3200 -R library as well as a -L library.
3201
3202 14. The debugging options of pcretest (-d on the command line or D on a
3203 pattern) showed incorrect output for anything following an extended class
3204 that contained multibyte characters and which was followed by a quantifier.
3205
3206 15. Added optional support for general category Unicode character properties
3207 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3208 support. It adds about 90K to the size of the library. The meanings of the
3209 inbuilt class escapes such as \d and \s have NOT been changed.
3210
3211 16. Updated pcredemo.c to include calls to free() to release the memory for the
3212 compiled pattern.
3213
3214 17. The generated file chartables.c was being created in the source directory
3215 instead of in the building directory. This caused the build to fail if the
3216 source directory was different from the building directory, and was
3217 read-only.
3218
3219 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3220 file. No doubt somebody will tell me if they don't make sense... Also added
3221 Dan Mooney's comments about building on OpenVMS.
3222
3223 19. Added support for partial matching via the PCRE_PARTIAL option for
3224 pcre_exec() and the \P data escape in pcretest.
3225
3226 20. Extended pcretest with 3 new pattern features:
3227
3228 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3229 write the compiled pattern to the file whose name is "rest-of-line".
3230 This is a straight binary dump of the data, with the saved pointer to
3231 the character tables forced to be NULL. The study data, if any, is
3232 written too. After writing, pcretest reads a new pattern.
3233
3234 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3235 compiled pattern from the given file. There must not be any
3236 occurrences of "<" in the file name (pretty unlikely); if there are,
3237 pcretest will instead treat the initial "<" as a pattern delimiter.
3238 After reading in the pattern, pcretest goes on to read data lines as
3239 usual.
3240
3241 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3242 and 16-bit fields in a compiled pattern, to simulate a pattern that
3243 was compiled on a host of opposite endianness.
3244
3245 21. The pcre-exec() function can now cope with patterns that were compiled on
3246 hosts of opposite endianness, with this restriction:
3247
3248 As for any compiled expression that is saved and used later, the tables
3249 pointer field cannot be preserved; the extra_data field in the arguments
3250 to pcre_exec() should be used to pass in a tables address if a value
3251 other than the default internal tables were used at compile time.
3252
3253 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3254 now diagnosed as an error. Previously, most of the time, a negative number
3255 would have been treated as zero, but if in addition "ovector" was passed as
3256 NULL, a crash could occur.
3257
3258 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3259 new versions from the libtool 1.5 distribution (the last one is a copy of
3260 a file called libtool.m4). This seems to have fixed the need to patch
3261 "configure" to support Darwin 1.3 (which I used to do). However, I still
3262 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3263 workstation).
3264
3265 24. Changed the PCRE licence to be the more standard "BSD" licence.
3266
3267
3268 Version 4.5 01-Dec-03
3269 ---------------------
3270
3271 1. There has been some re-arrangement of the code for the match() function so
3272 that it can be compiled in a version that does not call itself recursively.
3273 Instead, it keeps those local variables that need separate instances for
3274 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3275 needs to "recurse". Keeping track of where control must go is done by means
3276 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3277 hide most of the details from the main code, and operates only if
3278 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3279 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3280 operating.
3281
3282 To make it easier for callers to provide specially tailored get/free
3283 functions for this usage, two new functions, pcre_stack_malloc, and
3284 pcre_stack_free, are used. They are always called in strict stacking order,
3285 and the size of block requested is always the same.
3286
3287 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3288 PCRE has been compiled to use the stack or the heap for recursion. The
3289 -C option of pcretest uses this to show which version is compiled.
3290
3291 A new data escape \S, is added to pcretest; it causes the amounts of store
3292 obtained and freed by both kinds of malloc/free at match time to be added
3293 to the output.
3294
3295 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3296 what's available on my current Linux desktop machine.
3297
3298 3. When matching a UTF-8 string, the test for a valid string at the start has
3299 been extended. If start_offset is not zero, PCRE now checks that it points
3300 to a byte that is the start of a UTF-8 character. If not, it returns
3301 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3302 this is necessary because there may be backward assertions in the pattern.
3303 When matching the same subject several times, it may save resources to use
3304 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3305
3306 4. The code for checking the validity of UTF-8 strings has been tightened so
3307 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3308 containing "overlong sequences".
3309
3310 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3311 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3312 should have been "&", but it just so happened that all the cases this let
3313 through by mistake were picked up later in the function.
3314
3315 6. I had used a variable called "isblank" - this is a C99 function, causing
3316 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3317
3318 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3319 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3320 so the version is shown; (c) stop "make test" echoing "./RunTest".
3321
3322 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3323
3324 9. The prototype for memmove() for systems that don't have it was using
3325 size_t, but the inclusion of the header that defines size_t was later. I've
3326 moved the #includes for the C headers earlier to avoid this.
3327
3328 10. Added some adjustments to the code to make it easier to compiler on certain
3329 special systems:
3330
3331 (a) Some "const" qualifiers were missing.
3332 (b) Added the macro EXPORT before all exported functions; by default this
3333 is defined to be empty.
3334 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3335 that it reads its output file name as an argument instead of writing
3336 to the standard output and assuming this can be redirected.
3337
3338 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3339 class containing characters with values greater than 255, PCRE compilation
3340 went into a loop.
3341
3342 12. A recursive reference to a subpattern that was within another subpattern
3343 that had a minimum quantifier of zero caused PCRE to crash. For example,
3344 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3345 recursion. If the recursively-called subpattern itself had a zero repeat,
3346 that was OK.
3347
3348 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3349 buffer into which it was copied (for escape processing) was still set at
3350 1024, so long lines caused crashes.
3351
3352 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3353 "internal error: code overflow...". This applied to any character class
3354 that was followed by a possessive quantifier.
3355
3356 15. Modified the Makefile to add libpcre.la as a prerequisite for
3357 libpcreposix.la because I was told this is needed for a parallel build to
3358 work.
3359
3360 16. If a pattern that contained .* following optional items at the start was
3361 studied, the wrong optimizing data was generated, leading to matching
3362 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3363 matching string must start with a or b or c. The correct conclusion for
3364 this pattern is that a match can start with any character.
3365
3366
3367 Version 4.4 13-Aug-03
3368 ---------------------
3369
3370 1. In UTF-8 mode, a character class containing characters with values between
3371 127 and 255 was not handled correctly if the compiled pattern was studied.
3372 In fixing this, I have also improved the studying algorithm for such
3373 classes (slightly).
3374
3375 2. Three internal functions had redundant arguments passed to them. Removal
3376 might give a very teeny performance improvement.
3377
3378 3. Documentation bug: the value of the capture_top field in a callout is *one
3379 more than* the number of the hightest numbered captured substring.
3380
3381 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3382 in incorrectly linking with a previously installed version. They now link
3383 explicitly with libpcre.la.
3384
3385 5. configure.in no longer needs to recognize Cygwin specially.
3386
3387 6. A problem in pcre.in for Windows platforms is fixed.
3388
3389 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3390 pcretest, it used to include the size of the study block as part of its
3391 output. Unfortunately, the structure contains a field that has a different
3392 size on different hardware architectures. This meant that the tests that
3393 showed this size failed. As the block is currently always of a fixed size,
3394 this information isn't actually particularly useful in pcretest output, so
3395 I have just removed it.
3396
3397 8. Three pre-processor statements accidentally did not start in column 1.
3398 Sadly, there are *still* compilers around that complain, even though
3399 standard C has not required this for well over a decade. Sigh.
3400
3401 9. In pcretest, the code for checking callouts passed small integers in the
3402 callout_data field, which is a void * field. However, some picky compilers
3403 complained about the casts involved for this on 64-bit systems. Now
3404 pcretest passes the address of the small integer instead, which should get
3405 rid of the warnings.
3406
3407 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3408 both compile and run time, and gives an error if an invalid UTF-8 sequence
3409 is found. There is a option for disabling this check in cases where the
3410 string is known to be correct and/or the maximum performance is wanted.
3411
3412 11. In response to a bug report, I changed one line in Makefile.in from
3413
3414 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3415 to
3416 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3417
3418 to look similar to other lines, but I have no way of telling whether this
3419 is the right thing to do, as I do not use Windows. No doubt I'll get told
3420 if it's wrong...
3421
3422
3423 Version 4.3 21-May-03
3424 ---------------------
3425
3426 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3427 Makefile.
3428
3429 2. Some refactoring to improve the quality of the code:
3430
3431 (i) The utf8_table... variables are now declared "const".
3432
3433 (ii) The code for \cx, which used the "case flipping" table to upper case
3434 lower case letters, now just substracts 32. This is ASCII-specific,
3435 but the whole concept of \cx is ASCII-specific, so it seems
3436 reasonable.
3437
3438 (iii) PCRE was using its character types table to recognize decimal and
3439 hexadecimal digits in the pattern. This is silly, because it handles
3440 only 0-9, a-f, and A-F, but the character types table is locale-
3441 specific, which means strange things might happen. A private
3442 table is now used for this - though it costs 256 bytes, a table is
3443 much faster than multiple explicit tests. Of course, the standard
3444 character types table is still used for matching digits in subject
3445 strings against \d.
3446
3447 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3448 ending in _t are). So I've renamed it as ESC_tee.
3449
3450 3. The first argument for regexec() in the POSIX wrapper should have been
3451 defined as "const".
3452
3453 4. Changed pcretest to use malloc() for its buffers so that they can be
3454 Electric Fenced for debugging.
3455
3456 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3457 to read one or more bytes before the start of the subject string. Often this
3458 had no effect on PCRE's behaviour, but in some circumstances it could
3459 provoke a segmentation fault.
3460
3461 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3462 to try to read one or more bytes before the start of the subject string.
3463
3464 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3465 UTF-8 support could misbehave in various ways if the subject string
3466 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3467 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3468 back over UTF-8 characters.)
3469
3470
3471 Version 4.2 14-Apr-03
3472 ---------------------
3473
3474 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3475
3476 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3477 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3478 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3479 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3480 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3481 and BUILD_EXEEXT
3482 Note: automatic setting of the BUILD variables is not yet working
3483 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3484 compile-time but not at link-time
3485 [LINK]: use for linking executables only
3486 make different versions for Windows and non-Windows
3487 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3488 libraries
3489 [LINK_FOR_BUILD]: new variable
3490 [OBJEXT]: use throughout
3491 [EXEEXT]: use throughout
3492 <winshared>: new target
3493 <wininstall>: new target
3494 <dftables.o>: use native compiler
3495 <dftables>: use native linker
3496 <install>: handle Windows platform correctly
3497 <clean>: ditto
3498 <check>: ditto
3499 copy DLL to top builddir before testing
3500
3501 As part of these changes, -no-undefined was removed again. This was reported
3502 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3503 in any case.
3504
3505 3. Some tidies to get rid of compiler warnings:
3506
3507 . In the match_data structure, match_limit was an unsigned long int, whereas
3508 match_call_count was an int. I've made them both unsigned long ints.
3509
3510 . In pcretest the fact that a const uschar * doesn't automatically cast to
3511 a void * provoked a warning.
3512
3513 . Turning on some more compiler warnings threw up some "shadow" variables
3514 and a few more missing casts.
3515
3516 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3517 option, a class that contained a single character with a value between 128
3518 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3519
3520 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3521 option, a class that contained several characters, but with at least one
3522 whose value was between 128 and 255 caused PCRE to crash.
3523
3524
3525 Version 4.1 12-Mar-03
3526 ---------------------
3527
3528 1. Compiling with gcc -pedantic found a couple of places where casts were
3529 needed, and a string in dftables.c that was longer than standard compilers are
3530 required to support.
3531
3532 2. Compiling with Sun's compiler found a few more places where the code could
3533 be tidied up in order to avoid warnings.
3534
3535 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3536 first of these names is deprecated in the latest Autoconf in favour of the name
3537 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3538 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3539 analogy I have changed it to CFLAGS_FOR_BUILD.
3540
3541 4. Added -no-undefined to the linking command in the Makefile, because this is
3542 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3543 linking step for the pcreposix library.
3544
3545 5. PCRE was failing to diagnose the case of two named groups with the same
3546 name.
3547
3548 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3549 literal character that is needed in the subject for a match, and scans along to
3550 ensure that it is present before embarking on the full matching process. This
3551 saves time in cases of nested unlimited repeats that are never going to match.
3552 Problem: the scan can take a lot of time if the subject is very long (e.g.
3553 megabytes), thus penalizing straightforward matches. It is now done only if the
3554 amount of subject to be scanned is less than 1000 bytes.
3555
3556 7. A lesser problem with the same optimization is that it was recording the
3557 first character of an anchored pattern as "needed", thus provoking a search
3558 right along the subject, even when the first match of the pattern was going to
3559 fail. The "needed" character is now not set for anchored patterns, unless it
3560 follows something in the pattern that is of non-fixed length. Thus, it still
3561 fulfils its original purpose of finding quick non-matches in cases of nested
3562 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3563
3564
3565 Version 4.0 17-Feb-03
3566 ---------------------
3567
3568 1. If a comment in an extended regex that started immediately after a meta-item
3569 extended to the end of string, PCRE compiled incorrect data. This could lead to
3570 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3571
3572 2. Moved to autoconf 2.53 and libtool 1.4.2.
3573
3574 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3575 the special perltest8 script is no longer needed - all the tests can be run
3576 from a single perltest script.
3577
3578 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3579 by \s. It has now been removed in PCRE. This means it isn't recognized as
3580 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3581 class [:space:] *does* include VT, thereby creating a mess.
3582
3583 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3584 space and tab.
3585
3586 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3587 its new features into the main test script, reducing the number of scripts.
3588
3589 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3590 were backward compatible, and made the (?i) apply to the whole pattern, as if
3591 /i were given. Now it behaves more logically, and applies the option setting
3592 only to what follows. PCRE has been changed to follow suit. However, if it
3593 finds options settings right at the start of the pattern, it extracts them into
3594 the global options, as before. Thus, they show up in the info data.
3595
3596 8. Added support for the \Q...\E escape sequence. Characters in between are
3597 treated as literals. This is slightly different from Perl in that $ and @ are
3598 also handled as literals inside the quotes. In Perl, they will cause variable
3599 interpolation. Note the following examples:
3600
3601 Pattern PCRE matches Perl matches
3602
3603 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3604 \Qabc\$xyz\E abc\$xyz abc\$xyz
3605 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3606
3607 For compatibility with Perl, \Q...\E sequences are recognized inside character
3608 classes as well as outside them.
3609
3610 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3611 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3612 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3613 signed/unsigned warnings.
3614
3615 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3616 option for pcretest, so I've replaced it by a simple function that does just
3617 that job.
3618
3619 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3620 "pcregrep -".
3621
3622 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3623 Java package. This provides some syntactic sugar for simple cases of what my
3624 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3625 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3626 item, you can use this simplified notation. Note that only makes sense with
3627 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3628 greediness, whatever the setting of the PCRE_UNGREEDY option.
3629
3630 13. A change of greediness default within a pattern was not taking effect at
3631 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3632 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3633 was abstracted outside.
3634
3635 14. PCRE now supports the \G assertion. It is true when the current matching
3636 position is at the start point of the match. This differs from \A when the
3637 starting offset is non-zero. Used with the /g option of pcretest (or similar
3638 code), it works in the same way as it does for Perl's /g option. If all
3639 alternatives of a regex begin with \G, the expression is anchored to the start
3640 match position, and the "anchored" flag is set in the compiled expression.
3641
3642 15. Some bugs concerning the handling of certain option changes within patterns
3643 have been fixed. These applied to options other than (?ims). For example,
3644 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3645 been the other way round. Some of this was related to change 7 above.
3646
3647 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3648 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3649 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3650 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3651
3652 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3653 mode. Unlike ".", it always matches newline, whatever the setting of
3654 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3655 assertions. Perl allows it, but it doesn't (in general) work because it can't
3656 calculate the length of the lookbehind. At least, that's the case for Perl
3657 5.8.0 - I've been told they are going to document that it doesn't work in
3658 future.
3659
3660 18. Added an error diagnosis for escapes that PCRE does not support: these are
3661 \L, \l, \N, \P, \p, \U, \u, and \X.
3662
3663 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3664 reading past the end of the pattern in cases such as /[abcd/.
3665
3666 20. PCRE was getting more memory than necessary for patterns with classes that
3667 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3668
3669 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3670 compiling PCRE for use with Virtual Pascal.
3671
3672 22. Small fix to the Makefile to make it work properly if the build is done
3673 outside the source tree.
3674
3675 23. Added a new extension: a condition to go with recursion. If a conditional
3676 subpattern starts with (?(R) the "true" branch is used if recursion has
3677 happened, whereas the "false" branch is used only at the top level.
3678
3679 24. When there was a very long string of literal characters (over 255 bytes
3680 without UTF support, over 250 bytes with UTF support), the computation of how
3681 much memory was required could be incorrect, leading to segfaults or other
3682 strange effects.
3683
3684 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3685 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3686 there was a subsequent back reference to those brackets. This meant that, for
3687 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3688 possible to check for precisely this case. All we can do is abandon the
3689 optimization if .* occurs inside capturing brackets when there are any back
3690 references whatsoever. (See below for a better fix that came later.)
3691
3692 26. The handling of the optimization for finding the first character of a
3693 non-anchored pattern, and for finding a character that is required later in the
3694 match were failing in some cases. This didn't break the matching; it just
3695 failed to optimize when it could. The way this is done has been re-implemented.
3696
3697 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3698
3699 28. Added a new feature that provides some of the functionality that Perl
3700 provides with (?{...}). The facility is termed a "callout". The way it is done
3701 in PCRE is for the caller to provide an optional function, by setting
3702 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3703 global variable. By default it is unset, which disables all calling out. To get
3704 the function called, the regex must include (?C) at appropriate points. This
3705 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3706 This provides a means of identifying different callout points. When PCRE
3707 reaches such a point in the regex, if pcre_callout has been set, the external
3708 function is called. It is provided with data in a structure called
3709 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3710 matching continues; if it returns a non-zero value, the match at the current
3711 point fails. However, backtracking will occur if possible. [This was changed
3712 later and other features added - see item 49 below.]
3713
3714 29. pcretest is upgraded to test the callout functionality. It provides a
3715 callout function that displays information. By default, it shows the start of
3716 the match and the current position in the text. There are some new data escapes
3717 to vary what happens:
3718
3719 \C+ in addition, show current contents of captured substrings
3720 \C- do not supply a callout function
3721 \C!n return 1 when callout number n is reached
3722 \C!n!m return 1 when callout number n is reached for the mth time
3723
3724 30. If pcregrep was called with the -l option and just a single file name, it
3725 output "<stdin>" if a match was found, instead of the file name.
3726
3727 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3728 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3729 pcre_exec(). This saves a malloc/free per call. The default value of
3730 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3731 when configuring.
3732
3733 32. The default maximum size of a compiled pattern is 64K. There have been a
3734 few cases of people hitting this limit. The code now uses macros to handle the
3735 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3736 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3737 configuring. Tests 2 and 5 work only with 2-byte links because they output
3738 debugging information about compiled patterns.
3739
3740 33. Internal code re-arrangements:
3741
3742 (a) Moved the debugging function for printing out a compiled regex into
3743 its own source file (printint.c) and used #include to pull it into
3744 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3745 separate copies.
3746
3747 (b) Defined the list of op-code names for debugging as a macro in
3748 internal.h so that it is next to the definition of the opcodes.
3749
3750 (c) Defined a table of op-code lengths for simpler skipping along compiled
3751 code. This is again a macro in internal.h so that it is next to the
3752 definition of the opcodes.
3753
3754 34. Added support for recursive calls to individual subpatterns, along the
3755 lines of Robin Houston's patch (but implemented somewhat differently).
3756
3757 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3758 allow it to read and process whole directories in Win32. This code was
3759 contributed by Lionel Fourquaux; it has not been tested by me.
3760
3761 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3762 used to name a group. Names consist of alphanumerics and underscores, and must
3763 be unique. Back references use the syntax (?P=name) and recursive calls use
3764 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3765 numbers. The function pcre_fullinfo() can be used after compilation to extract
3766 a name/number map. There are three relevant calls:
3767
3768 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3769 PCRE_INFO_NAMECOUNT yields the number of entries
3770 PCRE_INFO_NAMETABLE yields a pointer to the map.
3771
3772 The map is a vector of fixed-size entries. The size of each entry depends on
3773 the length of the longest name used. The first two bytes of each entry are the
3774 group number, most significant byte first. There follows the corresponding
3775 name, zero terminated. The names are in alphabetical order.
3776
3777 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3778 case instead of 255. Making it the same both with and without UTF-8 support
3779 means that the same test output works with both.
3780
3781 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3782 calling malloc() with a zero argument.
3783
3784 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3785 optimization. I've improved things by keeping a bitmap of backreferences with
3786 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3787 fact referenced, the optimization can be applied. It is unlikely that a
3788 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3789 the match to follow \n) will appear inside brackets with a number greater than
3790 31, but if it does, any back reference > 31 suppresses the optimization.
3791
3792 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3793 of disabling numbered capturing parentheses. Any opening parenthesis that is
3794 not followed by ? behaves as if it were followed by ?: but named parentheses
3795 can still be used for capturing (and they will acquire numbers in the usual
3796 way).
3797
3798 41. Redesigned the return codes from the match() function into yes/no/error so
3799 that errors can be passed back from deep inside the nested calls. A malloc
3800 failure while inside a recursive subpattern call now causes the
3801 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3802
3803 42. It is now possible to set a limit on the number of times the match()
3804 function is called in a call to pcre_exec(). This facility makes it possible to
3805 limit the amount of recursion and backtracking, though not in a directly
3806 obvious way, because the match() function is used in a number of different
3807 circumstances. The count starts from zero for each position in the subject
3808 string (for non-anchored patterns). The default limit is, for compatibility, a
3809 large number, namely 10 000 000. You can change this in two ways:
3810
3811 (a) When configuring PCRE before making, you can use --with-match-limit=n
3812 to set a default value for the compiled library.
3813
3814 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3815 a different value is set. See 45 below.
3816
3817 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3818
3819 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3820 of things that can be changed at compile time. The first argument specifies
3821 what is wanted and the second points to where the information is to be placed.
3822 The current list of available information is:
3823
3824 PCRE_CONFIG_UTF8
3825
3826 The output is an integer that is set to one if UTF-8 support is available;
3827 otherwise it is set to zero.
3828
3829 PCRE_CONFIG_NEWLINE
3830
3831 The output is an integer that it set to the value of the code that is used for
3832 newline. It is either LF (10) or CR (13).
3833
3834 PCRE_CONFIG_LINK_SIZE
3835
3836 The output is an integer that contains the number of bytes used for internal
3837 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3838
3839 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3840
3841 The output is an integer that contains the threshold above which the POSIX
3842 interface uses malloc() for output vectors. See item 31 above.
3843
3844 PCRE_CONFIG_MATCH_LIMIT
3845
3846 The output is an unsigned integer that contains the default limit of the number
3847 of match() calls in a pcre_exec() execution. See 42 above.
3848
3849 44. pcretest has been upgraded by the addition of the -C option. This causes it
3850 to extract all the available output from the new pcre_config() function, and to
3851 output it. The program then exits immediately.
3852
3853 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3854 order to support additional features. One way would have been to define
3855 pcre_exec2() (for example) with extra arguments, but this would not have been
3856 extensible, and would also have required all calls to the original function to
3857 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3858 is used for passing in "extra" data from pcre_study().
3859
3860 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3861 contains the following fields:
3862
3863 flags a bitmap indicating which of the following fields are set
3864 study_data opaque data from pcre_study()
3865 match_limit a way of specifying a limit on match() calls for a specific
3866 call to pcre_exec()
3867 callout_data data for callouts (see 49 below)
3868
3869 The flag bits are also defined in pcre.h, and are
3870
3871 PCRE_EXTRA_STUDY_DATA
3872 PCRE_EXTRA_MATCH_LIMIT
3873 PCRE_EXTRA_CALLOUT_DATA
3874
3875 The pcre_study() function now returns one of these new pcre_extra blocks, with
3876 the actual study data pointed to by the study_data field, and the
3877 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3878 before. That is, this change is entirely upwards-compatible and requires no
3879 change to existing code.
3880
3881 If you want to pass in additional data to pcre_exec(), you can either place it
3882 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3883 block.
3884
3885 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3886 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3887 times with different match limits, until it finds the minimum value needed for
3888 pcre_exec() to complete. The value is then output. This can be instructive; for
3889 most simple matches the number is quite small, but for pathological cases it
3890 gets very large very quickly.
3891
3892 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3893 returns the size of the data block pointed to by the study_data field in a
3894 pcre_extra block, that is, the value that was passed as the argument to
3895 pcre_malloc() when PCRE was getting memory in which to place the information
3896 created by pcre_study(). The fourth argument should point to a size_t variable.
3897 pcretest has been extended so that this information is shown after a successful
3898 pcre_study() call when information about the compiled regex is being displayed.
3899
3900 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3901 because what follows is always an absolute path. (Later: it turns out that this
3902 is more than cosmetic for MinGW, because it doesn't like empty path
3903 components.)
3904
3905 49. Some changes have been made to the callout feature (see 28 above):
3906
3907 (i) A callout function now has three choices for what it returns:
3908
3909 0 => success, carry on matching
3910 > 0 => failure at this point, but backtrack if possible
3911 < 0 => serious error, return this value from pcre_exec()
3912
3913 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3914 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3915 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3916 use by callout functions. It will never be used by PCRE itself.
3917
3918 (ii) The pcre_extra structure (see 45 above) has a void * field called
3919 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3920 pcre_callout_block structure has a field of the same name. The contents of
3921 the field passed in the pcre_extra structure are passed to the callout
3922 function in the corresponding field in the callout block. This makes it
3923 easier to use the same callout-containing regex from multiple threads. For
3924 testing, the pcretest program has a new data escape
3925
3926 \C*n pass the number n (may be negative) as callout_data
3927
3928 If the callout function in pcretest receives a non-zero value as
3929 callout_data, it returns that value.
3930
3931 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3932 there were some redundant $(CFLAGS) in commands that are now specified as
3933 $(LINK), which already includes $(CFLAGS).
3934
3935 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3936 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3937 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3938 one-byte characters throughout. Note that case-insensitive matching applies
3939 only to characters whose values are less than 256. PCRE doesn't support the
3940 notion of cases for higher-valued characters.
3941
3942 (i) A character class whose characters are all within 0-255 is handled as
3943 a bit map, and the map is inverted for negative classes. Previously, a
3944 character > 255 always failed to match such a class; however it should
3945 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3946
3947 (ii) A negated character class with a single character < 255 is coded as
3948 "not this character" (OP_NOT). This wasn't working properly when the test
3949 character was multibyte, either singly or repeated.
3950
3951 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3952 mode, for example: \x{100}{2,3}.
3953
3954 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3955 singly or repeated) now correctly test multibyte characters. However,
3956 PCRE doesn't recognize any characters with values greater than 255 as
3957 digits, spaces, or word characters. Such characters always match \D, \S,
3958 and \W, and never match \d, \s, or \w.
3959
3960 (v) Classes may now contain characters and character ranges with values
3961 greater than 255. For example: [ab\x{100}-\x{400}].
3962
3963 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3964 PCRE in UTF-8 mode.
3965
3966 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3967 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3968 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3969 value.)
3970
3971 53. The single man page has become too large. I have therefore split it up into
3972 a number of separate man pages. These also give rise to individual HTML pages;
3973 these are now put in a separate directory, and there is an index.html page that
3974 lists them all. Some hyperlinking between the pages has been installed.
3975
3976 54. Added convenience functions for handling named capturing parentheses.
3977
3978 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3979 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3980 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3981 are faulted.
3982
3983 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3984 calling configure. These values are used when compiling the dftables.c program
3985 which is run to generate the source of the default character tables. They
3986 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3987 you will need to set these values.
3988
3989 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3990
3991
3992 Version 3.9 02-Jan-02
3993 ---------------------
3994
3995 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3996
3997 2. If --disable-static was given, the building process failed when trying to
3998 build pcretest and pcregrep. (For some reason it was using libtool to compile
3999 them, which is not right, as they aren't part of the library.)
4000
4001
4002 Version 3.8 18-Dec-01
4003 ---------------------
4004
4005 1. The experimental UTF-8 code was completely screwed up. It was packing the
4006 bytes in the wrong order. How dumb can you get?
4007
4008
4009 Version 3.7 29-Oct-01
4010 ---------------------
4011
4012 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4013 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4014 this didn't happen under Solaris 8, where I normally test things.
4015
4016 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4017 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4018 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4019 this fix an hour or so after the initial 3.7 release.)
4020
4021
4022 Version 3.6 23-Oct-01
4023 ---------------------
4024
4025 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4026 offsets passed as NULL with zero offset count.
4027
4028 2. The config.guess and config.sub files had not been updated when I moved to
4029 the latest autoconf.
4030
4031
4032 Version 3.5 15-Aug-01
4033 ---------------------
4034
4035 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4036 had been forgotten.
4037
4038 2. By using declared but undefined structures, we can avoid using "void"
4039 definitions in pcre.h while keeping the internal definitions of the structures
4040 private.
4041
4042 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4043 user point of view, this means that both static and shared libraries are built
4044 by default, but this can be individually controlled. More of the work of
4045 handling this static/shared cases is now inside libtool instead of PCRE's make
4046 file.
4047
4048 4. The pcretest utility is now installed along with pcregrep because it is
4049 useful for users (to test regexs) and by doing this, it automatically gets
4050 relinked by libtool. The documentation has been turned into a man page, so
4051 there are now .1, .txt, and .html versions in /doc.
4052
4053 5. Upgrades to pcregrep:
4054 (i) Added long-form option names like gnu grep.
4055 (ii) Added --help to list all options with an explanatory phrase.
4056 (iii) Added -r, --recursive to recurse into sub-directories.
4057 (iv) Added -f, --file to read patterns from a file.
4058
4059 6. pcre_exec() was referring to its "code" argument before testing that
4060 argument for NULL (and giving an error if it was NULL).
4061
4062 7. Upgraded Makefile.in to allow for compiling in a different directory from
4063 the source directory.
4064
4065 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4066 options bits, the pointer it was passed was to an int instead of to an unsigned
4067 long int. This mattered only on 64-bit systems.
4068
4069 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4070 generated) instead of pcre.in, which it its source. Also made the same change
4071 in several of the .c files.
4072
4073 10. A new release of gcc defines printf() as a macro, which broke pcretest
4074 because it had an ifdef in the middle of a string argument for printf(). Fixed
4075 by using separate calls to printf().
4076
4077 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4078 script, to force use of CR or LF instead of \n in the source. On non-Unix
4079 systems, the value can be set in config.h.
4080
4081 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4082 absolute limit. Changed the text of the error message to make this clear, and
4083 likewise updated the man page.
4084
4085 13. The limit of 99 on the number of capturing subpatterns has been removed.
4086 The new limit is 65535, which I hope will not be a "real" limit.
4087
4088
4089 Version 3.4 22-Aug-00
4090 ---------------------
4091
4092 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4093
4094 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4095
4096
4097 Version 3.3 01-Aug-00
4098 ---------------------
4099
4100 1. If an octal character was given, but the value was greater than \377, it
4101 was not getting masked to the least significant bits, as documented. This could
4102 lead to crashes in some systems.
4103
4104 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4105 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4106
4107 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4108 These just pass their arguments on to (pcre_free)(), but they are provided
4109 because some uses of PCRE bind it to non-C systems that can call its functions,
4110 but cannot call free() or pcre_free() directly.
4111
4112 4. Add "make test" as a synonym for "make check". Corrected some comments in
4113 the Makefile.
4114
4115 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4116 Makefile.
4117
4118 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4119 command called pgrep for grepping around the active processes.
4120
4121 7. Added the beginnings of support for UTF-8 character strings.
4122
4123 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4124 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4125 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4126 out for the ar command.)
4127
4128
4129 Version 3.2 12-May-00
4130 ---------------------
4131
4132 This is purely a bug fixing release.
4133
4134 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4135 of ZA. This was just one example of several cases that could provoke this bug,
4136 which was introduced by change 9 of version 2.00. The code for breaking
4137 infinite loops after an iteration that matches an empty string was't working
4138 correctly.
4139
4140 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4141 when matched against abbab (for example). After matching an empty string, it
4142 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4143 caused it to match further down the string than it should.
4144
4145 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4146 was there because it doesn't seem to be needed, and it causes trouble on some
4147 systems, as it is not a Standard C header. It has been removed.
4148
4149 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4150 were reported on the Macintosh. The changes were from
4151
4152 while ((c = *(++ptr)) != 0 && c != '\n');
4153 to
4154 while ((c = *(++ptr)) != 0 && c != '\n') ;
4155
4156 Totally extraordinary, but if that's what it takes...
4157
4158 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4159 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4160 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4161 assumes the way PCRE uses memmove() (always moving upwards).
4162
4163 6. PCRE is being used in one environment where strchr() is not available. There
4164 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4165 faster code anyway.
4166
4167
4168 Version 3.1 09-Feb-00
4169 ---------------------
4170
4171 The only change in this release is the fixing of some bugs in Makefile.in for
4172 the "install" target:
4173
4174 (1) It was failing to install pcreposix.h.
4175
4176 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4177
4178
4179 Version 3.0 01-Feb-00
4180 ---------------------
4181
4182 1. Add support for the /+ modifier to perltest (to output $` like it does in
4183 pcretest).
4184
4185 2. Add support for the /g modifier to perltest.
4186
4187 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4188 matches null strings.
4189
4190 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4191 pattern. Perl treats empty patterns specially - it reuses the most recent
4192 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4193 effect.
4194
4195 5. The POSIX interface was broken in that it was just handing over the POSIX
4196 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4197 required a bigger vector, with some working space on the end. This means that
4198 the POSIX wrapper now has to get and free some memory, and copy the results.
4199
4200 6. Added some simple autoconf support, placing the test data and the
4201 documentation in separate directories, re-organizing some of the
4202 information files, and making it build pcre-config (a GNU standard). Also added
4203 libtool support for building PCRE as a shared library, which is now the
4204 default.
4205
4206 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4207 09 are not valid octal constants. Single digits will be used for minor values
4208 less than 10.
4209
4210 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4211 existing programs that set these in the POSIX interface can use PCRE without
4212 modification.
4213
4214 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4215 return all that pcre_info() returns, plus additional data. The pcre_info()
4216 function is retained for compatibility, but is considered to be obsolete.
4217
4218 10. Added experimental recursion feature (?R) to handle one common case that
4219 Perl 5.6 will be able to do with (?p{...}).
4220
4221 11. Added support for POSIX character classes like [:alpha:], which Perl is
4222 adopting.
4223
4224
4225 Version 2.08 31-Aug-99
4226 ----------------------
4227
4228 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4229 trying to match at the startoffset position, but instead was moving forward to
4230 the next newline as if a previous match had failed.
4231
4232 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4233 and could get into a loop if a null string was matched other than at the start
4234 of the subject.
4235
4236 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4237 be distinguished at compile time, and for completeness also added PCRE_DATE.
4238
4239 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4240 in GnuWin32 environments.
4241
4242
4243 Version 2.07 29-Jul-99
4244 ----------------------
4245
4246 1. The documentation is now supplied in plain text form and HTML as well as in
4247 the form of man page sources.
4248
4249 2. C++ compilers don't like assigning (void *) values to other pointer types.
4250 In particular this affects malloc(). Although there is no problem in Standard
4251 C, I've put in casts to keep C++ compilers happy.
4252
4253 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4254 should be (const char *).
4255
4256 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4257 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4258 However, I haven't made this a standard facility. The documentation doesn't
4259 mention it, and the Makefile doesn't support it.
4260
4261 5. The Makefile now contains an "install" target, with editable destinations at
4262 the top of the file. The pcretest program is not installed.
4263
4264 6. pgrep -V now gives the PCRE version number and date.
4265
4266 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4267 causing the entire string to be ignored, instead of just the last character.
4268
4269 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4270 non-matching string, it can take a very, very long time, even for strings of
4271 quite modest length, because of the nested recursion. PCRE now does better in
4272 some of these cases. It does this by remembering the last required literal
4273 character in the pattern, and pre-searching the subject to ensure it is present
4274 before running the real match. In other words, it applies a heuristic to detect
4275 some types of certain failure quickly, and in the above example, if presented
4276 with a string that has no trailing " it gives "no match" very quickly.
4277
4278 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4279 other alternatives are tried instead.
4280
4281
4282 Version 2.06 09-Jun-99
4283 ----------------------
4284
4285 1. Change pcretest's output for amount of store used to show just the code
4286 space, because the remainder (the data block) varies in size between 32-bit and
4287 64-bit systems.
4288
4289 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4290 start matching at. This allows lookbehinds to work when searching for multiple
4291 occurrences in a string.
4292
4293 3. Added additional options to pcretest for testing multiple occurrences:
4294
4295 /+ outputs the rest of the string that follows a match
4296 /g loops for multiple occurrences, using the new startoffset argument
4297 /G loops for multiple occurrences by passing an incremented pointer
4298
4299 4. PCRE wasn't doing the "first character" optimization for patterns starting
4300 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4301 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4302 the letter 'x'. On long subject strings, this gives a significant speed-up.
4303
4304
4305 Version 2.05 21-Apr-99
4306 ----------------------
4307
4308 1. Changed the type of magic_number from int to long int so that it works
4309 properly on 16-bit systems.
4310
4311 2. Fixed a bug which caused patterns starting with .* not to work correctly
4312 when the subject string contained newline characters. PCRE was assuming
4313 anchoring for such patterns in all cases, which is not correct because .* will
4314 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4315 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4316 must be retried after every newline in the subject.
4317
4318
4319 Version 2.04 18-Feb-99
4320 ----------------------
4321
4322 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4323 computation of the store needed to hold the pattern was incorrect (too large).
4324 If such patterns were nested a few deep, this could multiply and become a real
4325 problem.
4326
4327 2. Added /M option to pcretest to show the memory requirement of a specific
4328 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4329
4330 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4331 compiled in such a way that the backtracking after subsequent failure was
4332 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4333 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4334
4335
4336 Version 2.03 02-Feb-99
4337 ----------------------
4338
4339 1. Fixed typo and small mistake in man page.
4340
4341 2. Added 4th condition (GPL supersedes if conflict) and created separate
4342 LICENCE file containing the conditions.
4343
4344 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4345 Perl, that is the internal \ allows the delimiter to be included in the
4346 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4347 the final delimiter, add \ to the end of the pattern (to test the error).
4348
4349 4. Added the convenience functions for extracting substrings after a successful
4350 match. Updated pcretest to make it able to test these functions.
4351
4352
4353 Version 2.02 14-Jan-99
4354 ----------------------
4355
4356 1. Initialized the working variables associated with each extraction so that
4357 their saving and restoring doesn't refer to uninitialized store.
4358
4359 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4360 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4361 fix the problem.
4362
4363 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4364 calls, and wasn't printing the correct value for compiling calls. Increased the
4365 default value of LOOPREPEAT, and the number of significant figures in the
4366 times.
4367
4368 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4369
4370 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4371 a building problem on Windows NT with a FAT file system.
4372
4373
4374 Version 2.01 21-Oct-98
4375 ----------------------
4376
4377 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4378 to character tables built by pcre_maketables() in the current locale. If NULL
4379 is passed, the default tables are used.
4380
4381
4382 Version 2.00 24-Sep-98
4383 ----------------------
4384
4385 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4386 it any more.
4387
4388 2. Allow quantification of (?>) groups, and make it work correctly.
4389
4390 3. The first character computation wasn't working for (?>) groups.
4391
4392 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4393 end of the subject) and add 5.005's \z, which really does match only at the
4394 very end of the subject.
4395
4396 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4397
4398 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4399 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4400 localized options. All options to pcre_study() were also removed.
4401
4402 7. Add other new features from 5.005:
4403
4404 $(?<= positive lookbehind
4405 $(?<! negative lookbehind
4406 (?imsx-imsx) added the unsetting capability
4407 such a setting is global if at outer level; local otherwise
4408 (?imsx-imsx:) non-capturing groups with option setting
4409 (?(cond)re|re) conditional pattern matching
4410
4411 A backreference to itself in a repeated group matches the previous
4412 captured string.
4413
4414 8. General tidying up of studying (both automatic and via "study")
4415 consequential on the addition of new assertions.
4416
4417 9. As in 5.005, unlimited repeated groups that could match an empty substring
4418 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4419 runtime if any iteration does actually match an empty substring.
4420
4421 10. Include the RunTest script in the distribution.
4422
4423 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4424 discrepancies, some of which were old and were also with respect to 5.004. They
4425 have now been fixed.
4426
4427
4428 Version 1.09 28-Apr-98
4429 ----------------------
4430
4431 1. A negated single character class followed by a quantifier with a minimum
4432 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4433 program crashes, or just wrong answers. This did not apply to negated classes
4434 containing more than one character, or to minima other than one.
4435
4436
4437 Version 1.08 27-Mar-98
4438 ----------------------
4439
4440 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4441
4442 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4443 latter must appear before anything that relies on it in the pattern.
4444
4445
4446 Version 1.07 16-Feb-98
4447 ----------------------
4448
4449 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4450 repeat of a potentially empty string).
4451
4452
4453 Version 1.06 23-Jan-98
4454 ----------------------
4455
4456 1. Added Markus Oberhumer's little patches for C++.
4457
4458 2. Literal strings longer than 255 characters were broken.
4459
4460
4461 Version 1.05 23-Dec-97
4462 ----------------------
4463
4464 1. Negated character classes containing more than one character were failing if
4465 PCRE_CASELESS was set at run time.
4466
4467
4468 Version 1.04 19-Dec-97
4469 ----------------------
4470
4471 1. Corrected the man page, where some "const" qualifiers had been omitted.
4472
4473 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4474 input syntax.
4475
4476 3. Fixed memory leak which occurred when a regex with back references was
4477 matched with an offsets vector that wasn't big enough. The temporary memory
4478 that is used in this case wasn't being freed if the match failed.
4479
4480 4. Tidied pcretest to ensure it frees memory that it gets.
4481
4482 5. Temporary memory was being obtained in the case where the passed offsets
4483 vector was exactly big enough.
4484
4485 6. Corrected definition of offsetof() from change 5 below.
4486
4487 7. I had screwed up change 6 below and broken the rules for the use of
4488 setjmp(). Now fixed.
4489
4490
4491 Version 1.03 18-Dec-97
4492 ----------------------
4493
4494 1. A erroneous regex with a missing opening parenthesis was correctly
4495 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4496 on some systems.
4497
4498 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4499 it was reported that one broken compiler failed on the former because "code" is
4500 also an independent variable.
4501
4502 3. The erroneous regex a[]b caused an array overrun reference.
4503
4504 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4505 fail on data ending with that character. (It was going on too far, and checking
4506 the next character, typically a binary zero.) This was specific to the
4507 optimized code for single-character negative classes.
4508
4509 5. Added a contributed patch from the TIN world which does the following:
4510
4511 + Add an undef for memmove, in case the the system defines a macro for it.
4512
4513 + Add a definition of offsetof(), in case there isn't one. (I don't know
4514 the reason behind this - offsetof() is part of the ANSI standard - but
4515 it does no harm).
4516
4517 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4518 most of the places where whitespace preceded '#'. I have given up and
4519 allowed the remaining 2 cases to be at the margin.
4520
4521 + Rename some variables in pcre to eliminate shadowing. This seems very
4522 pedantic, but does no harm, of course.
4523
4524 6. Moved the call to setjmp() into its own function, to get rid of warnings
4525 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4526
4527 7. Constructs such as \d{8,} were compiling into the equivalent of
4528 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4529 outcome, but in this particular case used more store than had been allocated,
4530 which caused the bug to be discovered because it threw up an internal error.
4531
4532 8. The debugging code in both pcre and pcretest for outputting the compiled
4533 form of a regex was going wrong in the case of back references followed by
4534 curly-bracketed repeats.
4535
4536
4537 Version 1.02 12-Dec-97
4538 ----------------------
4539
4540 1. Typos in pcre.3 and comments in the source fixed.
4541
4542 2. Applied a contributed patch to get rid of places where it used to remove
4543 'const' from variables, and fixed some signed/unsigned and uninitialized
4544 variable warnings.
4545
4546 3. Added the "runtest" target to Makefile.
4547
4548 4. Set default compiler flag to -O2 rather than just -O.
4549
4550
4551 Version 1.01 19-Nov-97
4552 ----------------------
4553
4554 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4555 like /([ab]*)*/, that is, for classes with more than one character in them.
4556
4557 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4558 as /((?>a*))*/ (a PCRE_EXTRA facility).
4559
4560
4561 Version 1.00 18-Nov-97
4562 ----------------------
4563
4564 1. Added compile-time macros to support systems such as SunOS4 which don't have
4565 memmove() or strerror() but have other things that can be used instead.
4566
4567 2. Arranged that "make clean" removes the executables.
4568
4569
4570 Version 0.99 27-Oct-97
4571 ----------------------
4572
4573 1. Fixed bug in code for optimizing classes with only one character. It was
4574 initializing a 32-byte map regardless, which could cause it to run off the end
4575 of the memory it had got.
4576
4577 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4578
4579
4580 Version 0.98 22-Oct-97
4581 ----------------------
4582
4583 1. Fixed bug in code for handling temporary memory usage when there are more
4584 back references than supplied space in the ovector. This could cause segfaults.
4585
4586
4587 Version 0.97 21-Oct-97
4588 ----------------------
4589
4590 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4591
4592 2. Optimized negated single characters not to use a bit map.
4593
4594 3. Brought error texts together as macro definitions; clarified some of them;
4595 fixed one that was wrong - it said "range out of order" when it meant "invalid
4596 escape sequence".
4597
4598 4. Changed some char * arguments to const char *.
4599
4600 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4601
4602 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4603 pcretest.
4604
4605
4606 Version 0.96 16-Oct-97
4607 ----------------------
4608
4609 1. Added a simple "pgrep" utility to the distribution.
4610
4611 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4612 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4613 where "ddd" means "one or more decimal digits".
4614
4615 3. Fixed serious bug. If a pattern had a back reference, but the call to
4616 pcre_exec() didn't supply a large enough ovector to record the related
4617 identifying subpattern, the match always failed. PCRE now remembers the number
4618 of the largest back reference, and gets some temporary memory in which to save
4619 the offsets during matching if necessary, in order to ensure that
4620 backreferences always work.
4621
4622 4. Increased the compatibility with Perl in a number of ways:
4623
4624 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4625 to request this handling. The option can be set at compile or exec time.
4626
4627 (b) $ matches before a terminating newline by default; an option
4628 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4629 mode). The option can be set at compile or exec time.
4630
4631 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4632 the same as Perl's. If the decimal number it represents is less than 10
4633 or there aren't that many previous left capturing parentheses, an octal
4634 escape is read. Inside a character class, it's always an octal escape,
4635 even if it is a single digit.
4636
4637 (d) An escaped but undefined alphabetic character is taken as a literal,
4638 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4639 escapes.
4640
4641 (e) {0} is now permitted. (The previous item is removed from the compiled
4642 pattern).
4643
4644 5. Changed all the names of code files so that the basic parts are no longer
4645 than 10 characters, and abolished the teeny "globals.c" file.
4646
4647 6. Changed the handling of character classes; they are now done with a 32-byte
4648 bit map always.
4649
4650 7. Added the -d and /D options to pcretest to make it possible to look at the
4651 internals of compilation without having to recompile pcre.
4652
4653
4654 Version 0.95 23-Sep-97
4655 ----------------------
4656
4657 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4658 \x20 at the start of a run of normal characters. These were being treated as
4659 real characters, instead of the source characters being re-checked.
4660
4661
4662 Version 0.94 18-Sep-97
4663 ----------------------
4664
4665 1. The functions are now thread-safe, with the caveat that the global variables
4666 containing pointers to malloc() and free() or alternative functions are the
4667 same for all threads.
4668
4669 2. Get pcre_study() to generate a bitmap of initial characters for non-
4670 anchored patterns when this is possible, and use it if passed to pcre_exec().
4671
4672
4673 Version 0.93 15-Sep-97
4674 ----------------------
4675
4676 1. /(b)|(:+)/ was computing an incorrect first character.
4677
4678 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4679 but not actually doing anything yet.
4680
4681 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4682 as Perl does (e.g. [-az] or [az-]).
4683
4684 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4685 all possible positions.
4686
4687 5. Split up into different modules to avoid including unneeded functions in a
4688 compiled binary. However, compile and exec are still in one module. The "study"
4689 function is split off.
4690
4691 6. The character tables are now in a separate module whose source is generated
4692 by an auxiliary program - but can then be edited by hand if required. There are
4693 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4694 toupper() in the code.
4695
4696 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4697 make them global. Abolish the function for setting them, as the caller can now
4698 set them directly.
4699
4700
4701 Version 0.92 11-Sep-97
4702 ----------------------
4703
4704 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4705 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4706
4707 2. Caseless matching was not working in character classes if the characters in
4708 the pattern were in upper case.
4709
4710 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4711
4712 4. Make PCRE_ANCHORED public and accept as a compile option.
4713
4714 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4715 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4716 pass them.
4717
4718 6. Give an error if bad option bits passed at compile or run time.
4719
4720 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4721 pcretest to cause it to pass that flag.
4722
4723 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4724 options, and the first character, if set.
4725
4726 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4727
4728
4729 Version 0.91 10-Sep-97
4730 ----------------------
4731
4732 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4733 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4734
4735 2. PCRE was looping on encountering an indefinitely repeated back reference to
4736 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4737 Perl does - treats the match as successful.
4738
4739 ****

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12