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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 953 - (show annotations)
Thu Mar 29 17:41:57 2012 UTC (3 years, 1 month ago) by zherczeg
File size: 228880 byte(s)
Error occurred while calculating annotation data.
Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler
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 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
79
80
81 Version 8.30 04-February-2012
82 -----------------------------
83
84 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
85 name is defined in ctype.h.
86
87 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
88 only in quite long subpatterns.
89
90 3. Removed the function pcre_info(), which has been obsolete and deprecated
91 since it was replaced by pcre_fullinfo() in February 2000.
92
93 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
94 match a (*MARK), and the match failed at the start of the subject, a
95 reference to memory before the start of the subject could occur. This bug
96 was introduced by fix 17 of release 8.21.
97
98 5. A reference to an unset group with zero minimum repetition was giving
99 totally wrong answers (in non-JavaScript-compatibility mode). For example,
100 /(another)?(\1?)test/ matched against "hello world test". This bug was
101 introduced in release 8.13.
102
103 6. Add support for 16-bit character strings (a large amount of work involving
104 many changes and refactorings).
105
106 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
107 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
108 from a file.
109
110 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
111 rounding is not applied in this particular case).
112
113 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
114 if they appear, or are escaped, in patterns.
115
116 10. Get rid of a number of -Wunused-but-set-variable warnings.
117
118 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
119 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
120 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
121 also returns the mark "x". This bug applied to capturing parentheses,
122 non-capturing parentheses, and atomic parentheses. It also applied to some
123 assertions.
124
125 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
126 information out of configure.ac instead of relying on pcre.h.generic, which
127 is not stored in the repository.
128
129 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
130 -lreadline.
131
132 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
133
134 15. Applied Graycode's patch to put the top-level frame on the stack rather
135 than the heap when not using the stack for recursion. This gives a
136 performance improvement in many cases when recursion is not deep.
137
138 16. Experimental code added to "pcretest -C" to output the stack frame size.
139
140
141 Version 8.21 12-Dec-2011
142 ------------------------
143
144 1. Updating the JIT compiler.
145
146 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
147 are added as well.
148
149 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
150 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
151 calling _pcre_jit_exec. Some extra comments are added.
152
153 4. (*MARK) settings inside atomic groups that do not contain any capturing
154 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
155 was introduced by change 18 for 8.20.
156
157 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
158 ECMA-262 standard.
159
160 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
161 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
162 This bug was probably introduced by change 9 of 8.13.
163
164 7. While fixing 6 above, I noticed that a number of other items were being
165 incorrectly rejected as "not fixed length". This arose partly because newer
166 opcodes had not been added to the fixed-length checking code. I have (a)
167 corrected the bug and added tests for these items, and (b) arranged for an
168 error to occur if an unknown opcode is encountered while checking for fixed
169 length instead of just assuming "not fixed length". The items that were
170 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
171 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
172 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
173
174 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
175 being incorrectly compiled and would have given unpredicatble results.
176
177 9. A possessively repeated subpattern with minimum repeat count greater than
178 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
179 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
180 the first (A) could occur when it should not.
181
182 10. Add a cast and remove a redundant test from the code.
183
184 11. JIT should use pcre_malloc/pcre_free for allocation.
185
186 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
187 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
188 is anything other than /usr, -L is still shown).
189
190 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
191
192 14. Perl does not support \N without a following name in a [] class; PCRE now
193 also gives an error.
194
195 15. If a forward reference was repeated with an upper limit of around 2000,
196 it caused the error "internal error: overran compiling workspace". The
197 maximum number of forward references (including repeats) was limited by the
198 internal workspace, and dependent on the LINK_SIZE. The code has been
199 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
200 the default depends on LINK_SIZE. There is a new upper limit (for safety)
201 of around 200,000 forward references. While doing this, I also speeded up
202 the filling in of repeated forward references.
203
204 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
205 incorrectly expecting the subject to contain another "a" after the start.
206
207 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
208 in the match, the SKIP should be ignored. This was not happening; instead
209 the SKIP was being treated as NOMATCH. For patterns such as
210 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
211 tested.
212
213 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
214 now much more compatible with Perl, in particular in cases where the result
215 is a non-match for a non-anchored pattern. For example, if
216 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
217 "m", where previously it did not return a name. A side effect of this
218 change is that for partial matches, the last encountered mark name is
219 returned, as for non matches. A number of tests that were previously not
220 Perl-compatible have been moved into the Perl-compatible test files. The
221 refactoring has had the pleasing side effect of removing one argument from
222 the match() function, thus reducing its stack requirements.
223
224 19. If the /S+ option was used in pcretest to study a pattern using JIT,
225 subsequent uses of /S (without +) incorrectly behaved like /S+.
226
227 21. Retrieve executable code size support for the JIT compiler and fixing
228 some warnings.
229
230 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
231 not work when the shorter character appeared right at the end of the
232 subject string.
233
234 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
235 systems.
236
237 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
238 output it when the /M option is used in pcretest.
239
240 25. The CheckMan script was not being included in the distribution. Also, added
241 an explicit "perl" to run Perl scripts from the PrepareRelease script
242 because this is reportedly needed in Windows.
243
244 26. If study data was being save in a file and studying had not found a set of
245 "starts with" bytes for the pattern, the data written to the file (though
246 never used) was taken from uninitialized memory and so caused valgrind to
247 complain.
248
249 27. Updated RunTest.bat as provided by Sheri Pierce.
250
251 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
252
253 29. Computation of memory usage for the table of capturing group names was
254 giving an unnecessarily large value.
255
256
257 Version 8.20 21-Oct-2011
258 ------------------------
259
260 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
261 a POSIX class. After further experiments with Perl, which convinced me that
262 Perl has bugs and confusions, a closing square bracket is no longer allowed
263 in a POSIX name. This bug also affected patterns with classes that started
264 with full stops.
265
266 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
267 captured substring, but while checking the failing first alternative,
268 substring 1 is temporarily captured. If the output vector supplied to
269 pcre_exec() was not big enough for this capture, the yield of the function
270 was still zero ("insufficient space for captured substrings"). This cannot
271 be totally fixed without adding another stack variable, which seems a lot
272 of expense for a edge case. However, I have improved the situation in cases
273 such as /(a)(b)x|abc/ matched against "abc", where the return code
274 indicates that fewer than the maximum number of slots in the ovector have
275 been set.
276
277 3. Related to (2) above: when there are more back references in a pattern than
278 slots in the output vector, pcre_exec() uses temporary memory during
279 matching, and copies in the captures as far as possible afterwards. It was
280 using the entire output vector, but this conflicts with the specification
281 that only 2/3 is used for passing back captured substrings. Now it uses
282 only the first 2/3, for compatibility. This is, of course, another edge
283 case.
284
285 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
286 main code base, and can be used by building with --enable-jit. When this is
287 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
288 runtime --no-jit option is given.
289
290 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
291 ovector, the return from the function was zero, implying that there were
292 other matches that did not fit. The correct "exactly full" value is now
293 returned.
294
295 6. If a subpattern that was called recursively or as a subroutine contained
296 (*PRUNE) or any other control that caused it to give a non-standard return,
297 invalid errors such as "Error -26 (nested recursion at the same subject
298 position)" or even infinite loops could occur.
299
300 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
301 computing the minimum length on reaching *ACCEPT, and so ended up with the
302 wrong value of 1 rather than 0. Further investigation indicates that
303 computing a minimum subject length in the presence of *ACCEPT is difficult
304 (think back references, subroutine calls), and so I have changed the code
305 so that no minimum is registered for a pattern that contains *ACCEPT.
306
307 8. If (*THEN) was present in the first (true) branch of a conditional group,
308 it was not handled as intended. [But see 16 below.]
309
310 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
311 Sheri Pierce.
312
313 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
314 the first byte in a match must be "a".
315
316 11. Change 17 for 8.13 increased the recursion depth for patterns like
317 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
318 pattern contains any instances of (*THEN). If it does not, the old
319 optimizations are restored. It would be nice to do this on a per-group
320 basis, but at the moment that is not feasible.
321
322 12. In some environments, the output of pcretest -C is CRLF terminated. This
323 broke RunTest's code that checks for the link size. A single white space
324 character after the value is now allowed for.
325
326 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
327 For "fr", it uses the Windows-specific input and output files.
328
329 14. If (*THEN) appeared in a group that was called recursively or as a
330 subroutine, it did not work as intended. [But see next item.]
331
332 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
333 pattern fragments (but not containing any | characters). If A and B are
334 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
335 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
336 D. In other words, Perl considers parentheses that do not contain any |
337 characters to be part of a surrounding alternative, whereas PCRE was
338 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
339 differently. PCRE now behaves in the same way as Perl, except in the case
340 of subroutine/recursion calls such as (?1) which have in any case always
341 been different (but PCRE had them first :-).
342
343 16. Related to 15 above: Perl does not treat the | in a conditional group as
344 creating alternatives. Such a group is treated in the same way as an
345 ordinary group without any | characters when processing (*THEN). PCRE has
346 been changed to match Perl's behaviour.
347
348 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
349 RunGrepTest script failed.
350
351 18. Change 22 for version 13 caused atomic groups to use more stack. This is
352 inevitable for groups that contain captures, but it can lead to a lot of
353 stack use in large patterns. The old behaviour has been restored for atomic
354 groups that do not contain any capturing parentheses.
355
356 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
357 suppress the check for a minimum subject length at run time. (If it was
358 given to pcre_exec() or pcre_dfa_exec() it did work.)
359
360 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
361 fail to work when decoding hex characters in data strings in EBCDIC
362 environments.
363
364 21. It appears that in at least one Mac OS environment, the isxdigit() function
365 is implemented as a macro that evaluates to its argument more than once,
366 contravening the C 90 Standard (I haven't checked a later standard). There
367 was an instance in pcretest which caused it to go wrong when processing
368 \x{...} escapes in subject strings. The has been rewritten to avoid using
369 things like p++ in the argument of isxdigit().
370
371
372 Version 8.13 16-Aug-2011
373 ------------------------
374
375 1. The Unicode data tables have been updated to Unicode 6.0.0.
376
377 2. Two minor typos in pcre_internal.h have been fixed.
378
379 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
380 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
381 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
382
383 4. There were a number of related bugs in the code for matching backrefences
384 caselessly in UTF-8 mode when codes for the characters concerned were
385 different numbers of bytes. For example, U+023A and U+2C65 are an upper
386 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
387 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
388 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
389 2-byte code at the end of the subject (it thought there wasn't enough data
390 left).
391
392 5. Comprehensive information about what went wrong is now returned by
393 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
394 as the output vector has at least 2 elements. The offset of the start of
395 the failing character and a reason code are placed in the vector.
396
397 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
398 now returned is for the first byte of the failing character, instead of the
399 last byte inspected. This is an incompatible change, but I hope it is small
400 enough not to be a problem. It makes the returned offset consistent with
401 pcre_exec() and pcre_dfa_exec().
402
403 7. pcretest now gives a text phrase as well as the error number when
404 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
405 failure, the offset and reason code are output.
406
407 8. When \R was used with a maximizing quantifier it failed to skip backwards
408 over a \r\n pair if the subsequent match failed. Instead, it just skipped
409 back over a single character (\n). This seems wrong (because it treated the
410 two characters as a single entity when going forwards), conflicts with the
411 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
412 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
413 has been changed.
414
415 9. Some internal refactoring has changed the processing so that the handling
416 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
417 time (the PCRE_DOTALL option was changed this way some time ago: version
418 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
419 which was always a bit of a fudge. It also means that there is one less
420 argument for the match() function, which reduces its stack requirements
421 slightly. This change also fixes an incompatibility with Perl: the pattern
422 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
423
424 10. More internal refactoring has drastically reduced the number of recursive
425 calls to match() for possessively repeated groups such as (abc)++ when
426 using pcre_exec().
427
428 11. While implementing 10, a number of bugs in the handling of groups were
429 discovered and fixed:
430
431 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
432 (a|)*(?1) gave a compile-time internal error.
433 ((a|)+)+ did not notice that the outer group could match an empty string.
434 (^a|^)+ was not marked as anchored.
435 (.*a|.*)+ was not marked as matching at start or after a newline.
436
437 12. Yet more internal refactoring has removed another argument from the match()
438 function. Special calls to this function are now indicated by setting a
439 value in a variable in the "match data" data block.
440
441 13. Be more explicit in pcre_study() instead of relying on "default" for
442 opcodes that mean there is no starting character; this means that when new
443 ones are added and accidentally left out of pcre_study(), testing should
444 pick them up.
445
446 14. The -s option of pcretest has been documented for ages as being an old
447 synonym of -m (show memory usage). I have changed it to mean "force study
448 for every regex", that is, assume /S for every regex. This is similar to -i
449 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
450 using it. It makes it easier to run collections of tests with and without
451 study enabled, and thereby test pcre_study() more easily. All the standard
452 tests are now run with and without -s (but some patterns can be marked as
453 "never study" - see 20 below).
454
455 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
456 restoration of the capturing data to the outer values was not happening
457 correctly.
458
459 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
460 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
461 pattern had matched an empty string, and so incorrectly returned a no
462 match.
463
464 17. There was optimizing code for the last branch of non-capturing parentheses,
465 and also for the obeyed branch of a conditional subexpression, which used
466 tail recursion to cut down on stack usage. Unfortunately, now that there is
467 the possibility of (*THEN) occurring in these branches, tail recursion is
468 no longer possible because the return has to be checked for (*THEN). These
469 two optimizations have therefore been removed. [But see 8.20/11 above.]
470
471 18. If a pattern containing \R was studied, it was assumed that \R always
472 matched two bytes, thus causing the minimum subject length to be
473 incorrectly computed because \R can also match just one byte.
474
475 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
476 was incorrectly computed.
477
478 20. If /S is present twice on a test pattern in pcretest input, it now
479 *disables* studying, thereby overriding the use of -s on the command line
480 (see 14 above). This is necessary for one or two tests to keep the output
481 identical in both cases.
482
483 21. When (*ACCEPT) was used in an assertion that matched an empty string and
484 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
485
486 22. When an atomic group that contained a capturing parenthesis was
487 successfully matched, but the branch in which it appeared failed, the
488 capturing was not being forgotten if a higher numbered group was later
489 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
490 group 1 to "a", when in fact it should be unset. This applied to multi-
491 branched capturing and non-capturing groups, repeated or not, and also to
492 positive assertions (capturing in negative assertions does not happen
493 in PCRE) and also to nested atomic groups.
494
495 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
496 subject after a captured substring, to make it easier to tell which of a
497 number of identical substrings has been captured.
498
499 24. The way atomic groups are processed by pcre_exec() has been changed so that
500 if they are repeated, backtracking one repetition now resets captured
501 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
502 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
503 "aaa". Previously, it would have been "a". As part of this code
504 refactoring, the way recursive calls are handled has also been changed.
505
506 25. If an assertion condition captured any substrings, they were not passed
507 back unless some other capturing happened later. For example, if
508 (?(?=(a))a) was matched against "a", no capturing was returned.
509
510 26. When studying a pattern that contained subroutine calls or assertions,
511 the code for finding the minimum length of a possible match was handling
512 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
513 group 1 called group 2 while simultaneously a separate group 2 called group
514 1). A stack overflow occurred in this case. I have fixed this by limiting
515 the recursion depth to 10.
516
517 27. Updated RunTest.bat in the distribution to the version supplied by Tom
518 Fortmann. This supports explicit test numbers on the command line, and has
519 argument validation and error reporting.
520
521 28. An instance of \X with an unlimited repeat could fail if at any point the
522 first character it looked at was a mark character.
523
524 29. Some minor code refactoring concerning Unicode properties and scripts
525 should reduce the stack requirement of match() slightly.
526
527 30. Added the '=' option to pcretest to check the setting of unused capturing
528 slots at the end of the pattern, which are documented as being -1, but are
529 not included in the return count.
530
531 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
532 compiled something random. Now it gives a compile-time error (as does
533 Perl).
534
535 32. A *MARK encountered during the processing of a positive assertion is now
536 recorded and passed back (compatible with Perl).
537
538 33. If --only-matching or --colour was set on a pcregrep call whose pattern
539 had alternative anchored branches, the search for a second match in a line
540 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
541 matched the line "0102" twice. The same bug affected patterns that started
542 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
543 twice.
544
545 34. Previously, PCRE did not allow quantification of assertions. However, Perl
546 does, and because of capturing effects, quantifying parenthesized
547 assertions may at times be useful. Quantifiers are now allowed for
548 parenthesized assertions.
549
550 35. A minor code tidy in pcre_compile() when checking options for \R usage.
551
552 36. \g was being checked for fancy things in a character class, when it should
553 just be a literal "g".
554
555 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
556 appearance of a nested POSIX class supersedes an apparent external class.
557 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
558 unescaped square brackets may also appear as part of class names. For
559 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
560 more like Perl. (But see 8.20/1 above.)
561
562 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
563 was because it thought it was \N{name}, which is not supported).
564
565 39. Add minix to OS list not supporting the -S option in pcretest.
566
567 40. PCRE tries to detect cases of infinite recursion at compile time, but it
568 cannot analyze patterns in sufficient detail to catch mutual recursions
569 such as ((?1))((?2)). There is now a runtime test that gives an error if a
570 subgroup is called recursively as a subpattern for a second time at the
571 same position in the subject string. In previous releases this might have
572 been caught by the recursion limit, or it might have run out of stack.
573
574 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
575 happen only once. PCRE was, however incorrectly giving a compile time error
576 "recursive call could loop indefinitely" because it cannot analyze the
577 pattern in sufficient detail. The compile time test no longer happens when
578 PCRE is compiling a conditional subpattern, but actual runaway loops are
579 now caught at runtime (see 40 above).
580
581 42. It seems that Perl allows any characters other than a closing parenthesis
582 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
583 has been changed to be the same.
584
585 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
586 as not to get warnings when autogen.sh is called. Also changed
587 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
588
589 44. To help people who use pcregrep to scan files containing exceedingly long
590 lines, the following changes have been made:
591
592 (a) The default value of the buffer size parameter has been increased from
593 8K to 20K. (The actual buffer used is three times this size.)
594
595 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
596 PCRE is built.
597
598 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
599 to be set at run time.
600
601 (d) Numerical values in pcregrep options can be followed by K or M, for
602 example --buffer-size=50K.
603
604 (e) If a line being scanned overflows pcregrep's buffer, an error is now
605 given and the return code is set to 2.
606
607 45. Add a pointer to the latest mark to the callout data block.
608
609 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
610 partial match of an empty string instead of no match. This was specific to
611 the use of ".".
612
613 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
614 complete match instead of a partial match. This bug was dependent on both
615 the PCRE_UTF8 and PCRE_DOTALL options being set.
616
617 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
618 starting byte set, because \b was not being ignored.
619
620
621 Version 8.12 15-Jan-2011
622 ------------------------
623
624 1. Fixed some typos in the markup of the man pages, and wrote a script that
625 checks for such things as part of the documentation building process.
626
627 2. On a big-endian 64-bit system, pcregrep did not correctly process the
628 --match-limit and --recursion-limit options (added for 8.11). In
629 particular, this made one of the standard tests fail. (The integer value
630 went into the wrong half of a long int.)
631
632 3. If the --colour option was given to pcregrep with -v (invert match), it
633 did strange things, either producing crazy output, or crashing. It should,
634 of course, ignore a request for colour when reporting lines that do not
635 match.
636
637 4. Another pcregrep bug caused similar problems if --colour was specified with
638 -M (multiline) and the pattern match finished with a line ending.
639
640 5. In pcregrep, when a pattern that ended with a literal newline sequence was
641 matched in multiline mode, the following line was shown as part of the
642 match. This seems wrong, so I have changed it.
643
644 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
645 the check for further matches in the same line (so they could be coloured)
646 to overrun the end of the current line. If another match was found, it was
647 incorrectly shown (and then shown again when found in the next line).
648
649 7. If pcregrep was compiled under Windows, there was a reference to the
650 function pcregrep_exit() before it was defined. I am assuming this was
651 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
652 reported by a user. I've moved the definition above the reference.
653
654
655 Version 8.11 10-Dec-2010
656 ------------------------
657
658 1. (*THEN) was not working properly if there were untried alternatives prior
659 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
660 backtracked to try for "b" instead of moving to the next alternative branch
661 at the same level (in this case, to look for "c"). The Perl documentation
662 is clear that when (*THEN) is backtracked onto, it goes to the "next
663 alternative in the innermost enclosing group".
664
665 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
666 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
667 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
668 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
669 (*THEN).
670
671 3. If \s appeared in a character class, it removed the VT character from
672 the class, even if it had been included by some previous item, for example
673 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
674 of \s, but is part of the POSIX "space" class.)
675
676 4. A partial match never returns an empty string (because you can always
677 match an empty string at the end of the subject); however the checking for
678 an empty string was starting at the "start of match" point. This has been
679 changed to the "earliest inspected character" point, because the returned
680 data for a partial match starts at this character. This means that, for
681 example, /(?<=abc)def/ gives a partial match for the subject "abc"
682 (previously it gave "no match").
683
684 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
685 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
686 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
687 has an implication that the given string is incomplete (because a partial
688 match is preferred over a full match). For this reason, these items now
689 give a partial match in this situation. [Aside: previously, the one case
690 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
691 match rather than a full match, which was wrong by the old rules, but is
692 now correct.]
693
694 6. There was a bug in the handling of #-introduced comments, recognized when
695 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
696 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
697 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
698 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
699 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
700 places in pcre_compile().
701
702 7. Related to (6) above, when pcre_compile() was skipping #-introduced
703 comments when looking ahead for named forward references to subpatterns,
704 the only newline sequence it recognized was NL. It now handles newlines
705 according to the set newline convention.
706
707 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
708 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
709 cater for a lack of strerror(). These oversights have been fixed.
710
711 9. Added --match-limit and --recursion-limit to pcregrep.
712
713 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
714
715 11. When the -o option was used, pcregrep was setting a return code of 1, even
716 when matches were found, and --line-buffered was not being honoured.
717
718 12. Added an optional parentheses number to the -o and --only-matching options
719 of pcregrep.
720
721 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
722 can match an empty string. The code to do it in pcretest and pcredemo
723 needed fixing:
724
725 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
726 only one byte after an empty string match just before CRLF (this case
727 just got forgotten; "any" and "anycrlf" were OK).
728
729 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
730 mode when an empty string match preceded an ASCII character followed by
731 a non-ASCII character. (The code for advancing by one character rather
732 than one byte was nonsense.)
733
734 (c) The pcredemo.c sample program did not have any code at all to handle
735 the cases when CRLF is a valid newline sequence.
736
737 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
738 as a starting offset was within the subject string. There is now a new
739 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
740 negative or greater than the length of the string. In order to test this,
741 pcretest is extended to allow the setting of negative starting offsets.
742
743 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
744 starting offset points to the beginning of a UTF-8 character was
745 unnecessarily clumsy. I tidied it up.
746
747 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
748 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
749
750 17. Nobody had reported that the --include_dir option, which was added in
751 release 7.7 should have been called --include-dir (hyphen, not underscore)
752 for compatibility with GNU grep. I have changed it to --include-dir, but
753 left --include_dir as an undocumented synonym, and the same for
754 --exclude-dir, though that is not available in GNU grep, at least as of
755 release 2.5.4.
756
757 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
758 characters from a string of bytes) have been redefined so as not to use
759 loops, in order to improve performance in some environments. At the same
760 time, I abstracted some of the common code into auxiliary macros to save
761 repetition (this should not affect the compiled code).
762
763 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
764 compile-time error is now given if \c is not followed by an ASCII
765 character, that is, a byte less than 128. (In EBCDIC mode, the code is
766 different, and any byte value is allowed.)
767
768 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
769 START_OPTIMIZE option, which is now allowed at compile time - but just
770 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
771 to pcregrep and other applications that have no direct access to PCRE
772 options. The new /Y option in pcretest sets this option when calling
773 pcre_compile().
774
775 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
776 back references. Groups containing recursive back references were forced to
777 be atomic by that change, but in the case of named groups, the amount of
778 memory required was incorrectly computed, leading to "Failed: internal
779 error: code overflow". This has been fixed.
780
781 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
782 pcretest.c, to avoid build problems in some Borland environments.
783
784
785 Version 8.10 25-Jun-2010
786 ------------------------
787
788 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
789 THEN.
790
791 2. (*ACCEPT) was not working when inside an atomic group.
792
793 3. Inside a character class, \B is treated as a literal by default, but
794 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
795 causes the error). The code is unchanged, but I tidied the documentation.
796
797 4. Inside a character class, PCRE always treated \R and \X as literals,
798 whereas Perl faults them if its -w option is set. I have changed PCRE so
799 that it faults them when PCRE_EXTRA is set.
800
801 5. Added support for \N, which always matches any character other than
802 newline. (It is the same as "." when PCRE_DOTALL is not set.)
803
804 6. When compiling pcregrep with newer versions of gcc which may have
805 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
806 declared with attribute warn_unused_result" were given. Just casting the
807 result to (void) does not stop the warnings; a more elaborate fudge is
808 needed. I've used a macro to implement this.
809
810 7. Minor change to pcretest.c to avoid a compiler warning.
811
812 8. Added four artifical Unicode properties to help with an option to make
813 \s etc use properties (see next item). The new properties are: Xan
814 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
815
816 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
817 use Unicode properties. (*UCP) at the start of a pattern can be used to set
818 this option. Modified pcretest to add /W to test this facility. Added
819 REG_UCP to make it available via the POSIX interface.
820
821 10. Added --line-buffered to pcregrep.
822
823 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
824 studied, and the match started with a letter with a code point greater than
825 127 whose first byte was different to the first byte of the other case of
826 the letter, the other case of this starting letter was not recognized
827 (#976).
828
829 12. If a pattern that was studied started with a repeated Unicode property
830 test, for example, \p{Nd}+, there was the theoretical possibility of
831 setting up an incorrect bitmap of starting bytes, but fortunately it could
832 not have actually happened in practice until change 8 above was made (it
833 added property types that matched character-matching opcodes).
834
835 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
836 possible starting bytes for non-anchored patterns.
837
838 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
839 \R, and also a number of cases that involve Unicode properties, both
840 explicit and implicit when PCRE_UCP is set.
841
842 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
843 input, it could crash or give wrong results if characters with values
844 greater than 0xc0 were present in the subject string. (Detail: it assumed
845 UTF-8 input when processing these items.)
846
847 16. Added a lot of (int) casts to avoid compiler warnings in systems where
848 size_t is 64-bit (#991).
849
850 17. Added a check for running out of memory when PCRE is compiled with
851 --disable-stack-for-recursion (#990).
852
853 18. If the last data line in a file for pcretest does not have a newline on
854 the end, a newline was missing in the output.
855
856 19. The default pcre_chartables.c file recognizes only ASCII characters (values
857 less than 128) in its various bitmaps. However, there is a facility for
858 generating tables according to the current locale when PCRE is compiled. It
859 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
860 characters, are recognized by isspace() and therefore were getting set in
861 these tables, and indeed these tables seem to approximate to ISO 8859. This
862 caused a problem in UTF-8 mode when pcre_study() was used to create a list
863 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
864 which of course cannot start UTF-8 characters. I have changed the code so
865 that only real ASCII characters (less than 128) and the correct starting
866 bytes for UTF-8 encodings are set for characters greater than 127 when in
867 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
868 altogether.)
869
870 20. Added the /T option to pcretest so as to be able to run tests with non-
871 standard character tables, thus making it possible to include the tests
872 used for 19 above in the standard set of tests.
873
874 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
875 reference to a subpattern the other side of a comment that contains an
876 opening parenthesis caused either an internal compiling error, or a
877 reference to the wrong subpattern.
878
879
880 Version 8.02 19-Mar-2010
881 ------------------------
882
883 1. The Unicode data tables have been updated to Unicode 5.2.0.
884
885 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
886 configured.
887
888 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
889 original author of that file, following a query about its status.
890
891 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
892 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
893
894 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
895 quantifier applied to a forward-referencing subroutine call, could compile
896 incorrect code or give the error "internal error: previously-checked
897 referenced subpattern not found".
898
899 6. Both MS Visual Studio and Symbian OS have problems with initializing
900 variables to point to external functions. For these systems, therefore,
901 pcre_malloc etc. are now initialized to local functions that call the
902 relevant global functions.
903
904 7. There were two entries missing in the vectors called coptable and poptable
905 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
906 I've fixed the data, and added a kludgy way of testing at compile time that
907 the lengths are correct (equal to the number of opcodes).
908
909 8. Following on from 7, I added a similar kludge to check the length of the
910 eint vector in pcreposix.c.
911
912 9. Error texts for pcre_compile() are held as one long string to avoid too
913 much relocation at load time. To find a text, the string is searched,
914 counting zeros. There was no check for running off the end of the string,
915 which could happen if a new error number was added without updating the
916 string.
917
918 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
919
920 11. \K was not working if it appeared in an atomic group or in a group that
921 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
922 \K is "not well defined" if used in an assertion. PCRE now accepts it if
923 the assertion is positive, but not if it is negative.
924
925 12. Change 11 fortuitously reduced the size of the stack frame used in the
926 "match()" function of pcre_exec.c by one pointer. Forthcoming
927 implementation of support for (*MARK) will need an extra pointer on the
928 stack; I have reserved it now, so that the stack frame size does not
929 decrease.
930
931 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
932 item in branch that calls a recursion is a subroutine call - as in the
933 second branch in the above example - was incorrectly given the compile-
934 time error "recursive call could loop indefinitely" because pcre_compile()
935 was not correctly checking the subroutine for matching a non-empty string.
936
937 14. The checks for overrunning compiling workspace could trigger after an
938 overrun had occurred. This is a "should never occur" error, but it can be
939 triggered by pathological patterns such as hundreds of nested parentheses.
940 The checks now trigger 100 bytes before the end of the workspace.
941
942 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
943
944
945 Version 8.01 19-Jan-2010
946 ------------------------
947
948 1. If a pattern contained a conditional subpattern with only one branch (in
949 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
950 computed the wrong minimum data length (which is of course zero for such
951 subpatterns). This could cause incorrect "no match" results.
952
953 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
954 the pattern is reset in the first branch, pcre_compile() failed with
955 "internal error: code overflow at offset...". This happened only when
956 the reset was to the original external option setting. (An optimization
957 abstracts leading options settings into an external setting, which was the
958 cause of this.)
959
960 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
961 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
962 assertion pattern did not match (meaning that the assertion was true), it
963 was incorrectly treated as false if the SKIP had been reached during the
964 matching. This also applied to assertions used as conditions.
965
966 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
967 assertion subpattern, including such a pattern used as a condition,
968 unpredictable results occurred, instead of the error return
969 PCRE_ERROR_DFA_UITEM.
970
971 5. The C++ GlobalReplace function was not working like Perl for the special
972 situation when an empty string is matched. It now does the fancy magic
973 stuff that is necessary.
974
975 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
976 removed. (These were left over from very, very early versions of PCRE.)
977
978 7. Some cosmetic changes to the code to make life easier when compiling it
979 as part of something else:
980
981 (a) Change DEBUG to PCRE_DEBUG.
982
983 (b) In pcre_compile(), rename the member of the "branch_chain" structure
984 called "current" as "current_branch", to prevent a collision with the
985 Linux macro when compiled as a kernel module.
986
987 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
988 prevent a collision with the Linux macro when compiled as a kernel
989 module.
990
991 8. In pcre_compile() there are some checks for integer overflows that used to
992 cast potentially large values to (double). This has been changed to that
993 when building, a check for int64_t is made, and if it is found, it is used
994 instead, thus avoiding the use of floating point arithmetic. (There is no
995 other use of FP in PCRE.) If int64_t is not found, the fallback is to
996 double.
997
998 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
999 2005 (difference between two addresses compared to an unsigned value).
1000
1001 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1002 custom one, because of the following reported problem in Windows:
1003
1004 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1005 under Win32.
1006 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1007 therefore missing the function definition.
1008 - The compiler thus generates a "C" signature for the test function.
1009 - The linker fails to find the "C" function.
1010 - PCRE fails to configure if asked to do so against libbz2.
1011
1012 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1013 messages were output:
1014
1015 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1016 rerunning libtoolize, to keep the correct libtool macros in-tree.
1017 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1018
1019 I have done both of these things.
1020
1021 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1022 most of the time, it *can* run out if it is given a pattern that contains a
1023 runaway infinite recursion. I updated the discussion in the pcrestack man
1024 page.
1025
1026 13. Now that we have gone to the x.xx style of version numbers, the minor
1027 version may start with zero. Using 08 or 09 is a bad idea because users
1028 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1029 interpreted as invalid octal numbers. I've updated the previous comment in
1030 configure.ac, and also added a check that gives an error if 08 or 09 are
1031 used.
1032
1033 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1034 causing partial matching to fail when the end of the subject matched \W
1035 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1036
1037 15. There were some discrepancies between the declarations in pcre_internal.h
1038 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1039 their definitions. The declarations used "const uschar *" and the
1040 definitions used USPTR. Even though USPTR is normally defined as "const
1041 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1042 reported that: "This difference in casting confuses some C++ compilers, for
1043 example, SunCC recognizes above declarations as different functions and
1044 generates broken code for hbpcre." I have changed the declarations to use
1045 USPTR.
1046
1047 16. GNU libtool is named differently on some systems. The autogen.sh script now
1048 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1049 (FreeBSD).
1050
1051 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1052 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1053 comment: "Figure out how to create a longlong from a string: strtoll and
1054 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1055 instance, but it only takes 2 args instead of 3!"
1056
1057 18. A subtle bug concerned with back references has been fixed by a change of
1058 specification, with a corresponding code fix. A pattern such as
1059 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1060 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1061 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1062 same bug. Such groups have to be quantified to be useful, or contained
1063 inside another quantified group. (If there's no repetition, the reference
1064 can never match.) The problem arises because, having left the group and
1065 moved on to the rest of the pattern, a later failure that backtracks into
1066 the group uses the captured value from the final iteration of the group
1067 rather than the correct earlier one. I have fixed this in PCRE by forcing
1068 any group that contains a reference to itself to be an atomic group; that
1069 is, there cannot be any backtracking into it once it has completed. This is
1070 similar to recursive and subroutine calls.
1071
1072
1073 Version 8.00 19-Oct-09
1074 ----------------------
1075
1076 1. The table for translating pcre_compile() error codes into POSIX error codes
1077 was out-of-date, and there was no check on the pcre_compile() error code
1078 being within the table. This could lead to an OK return being given in
1079 error.
1080
1081 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1082 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1083 in a Windows environment.
1084
1085 3. The pcregrep --count option prints the count for each file even when it is
1086 zero, as does GNU grep. However, pcregrep was also printing all files when
1087 --files-with-matches was added. Now, when both options are given, it prints
1088 counts only for those files that have at least one match. (GNU grep just
1089 prints the file name in this circumstance, but including the count seems
1090 more useful - otherwise, why use --count?) Also ensured that the
1091 combination -clh just lists non-zero counts, with no names.
1092
1093 4. The long form of the pcregrep -F option was incorrectly implemented as
1094 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1095 but it seems right to fix it, and I didn't think it was worth preserving
1096 the old behaviour.
1097
1098 5. The command line items --regex=pattern and --regexp=pattern were not
1099 recognized by pcregrep, which required --regex pattern or --regexp pattern
1100 (with a space rather than an '='). The man page documented the '=' forms,
1101 which are compatible with GNU grep; these now work.
1102
1103 6. No libpcreposix.pc file was created for pkg-config; there was just
1104 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1105
1106 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1107 when UCP support is not needed, by modifying the Python script that
1108 generates it from Unicode data files. This should not matter if the module
1109 is correctly used as a library, but I received one complaint about 50K of
1110 unwanted data. My guess is that the person linked everything into his
1111 program rather than using a library. Anyway, it does no harm.
1112
1113 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1114 was a minimum greater than 1 for a wide character in a possessive
1115 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1116 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1117 character. Chaos in the form of incorrect output or a compiling loop could
1118 result.
1119
1120 9. The restrictions on what a pattern can contain when partial matching is
1121 requested for pcre_exec() have been removed. All patterns can now be
1122 partially matched by this function. In addition, if there are at least two
1123 slots in the offset vector, the offset of the earliest inspected character
1124 for the match and the offset of the end of the subject are set in them when
1125 PCRE_ERROR_PARTIAL is returned.
1126
1127 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1128 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1129 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1130 and may be more useful for multi-segment matching.
1131
1132 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1133 used to be given if ever the end of the subject was reached; now it is
1134 given only if matching could not proceed because another character was
1135 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1136 string "Z", which now yields "no match" instead of "partial match". In the
1137 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1138 final character ended with (*FAIL).
1139
1140 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1141 if the pattern had a "must contain" character that was already found in the
1142 earlier partial match, unless partial matching was again requested. For
1143 example, with the pattern /dog.(body)?/, the "must contain" character is
1144 "g". If the first part-match was for the string "dog", restarting with
1145 "sbody" failed. This bug has been fixed.
1146
1147 13. The string returned by pcre_dfa_exec() after a partial match has been
1148 changed so that it starts at the first inspected character rather than the
1149 first character of the match. This makes a difference only if the pattern
1150 starts with a lookbehind assertion or \b or \B (\K is not supported by
1151 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1152 matching functions compatible, and I think it's the right thing to do.
1153
1154 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1155 so that the demonstration program is easily available in environments where
1156 PCRE has not been installed from source.
1157
1158 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1159 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1160 library.
1161
1162 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1163 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1164 is not the first non-POSIX option to be added. Clearly some people find
1165 these options useful.
1166
1167 17. If a caller to the POSIX matching function regexec() passes a non-zero
1168 value for nmatch with a NULL value for pmatch, the value of
1169 nmatch is forced to zero.
1170
1171 18. RunGrepTest did not have a test for the availability of the -u option of
1172 the diff command, as RunTest does. It now checks in the same way as
1173 RunTest, and also checks for the -b option.
1174
1175 19. If an odd number of negated classes containing just a single character
1176 interposed, within parentheses, between a forward reference to a named
1177 subpattern and the definition of the subpattern, compilation crashed with
1178 an internal error, complaining that it could not find the referenced
1179 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1180 [The bug was that it was starting one character too far in when skipping
1181 over the character class, thus treating the ] as data rather than
1182 terminating the class. This meant it could skip too much.]
1183
1184 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1185 /g option in pcretest when the pattern contains \K, which makes it possible
1186 to have an empty string match not at the start, even when the pattern is
1187 anchored. Updated pcretest and pcredemo to use this option.
1188
1189 21. If the maximum number of capturing subpatterns in a recursion was greater
1190 than the maximum at the outer level, the higher number was returned, but
1191 with unset values at the outer level. The correct (outer level) value is
1192 now given.
1193
1194 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1195 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1196 make it do so. The string so far is captured, making this feature
1197 compatible with Perl.
1198
1199 23. The tests have been re-organized, adding tests 11 and 12, to make it
1200 possible to check the Perl 5.10 features against Perl 5.10.
1201
1202 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1203 pattern matches a fixed length string. PCRE did not allow this; now it
1204 does. Neither allows recursion.
1205
1206 25. I finally figured out how to implement a request to provide the minimum
1207 length of subject string that was needed in order to match a given pattern.
1208 (It was back references and recursion that I had previously got hung up
1209 on.) This code has now been added to pcre_study(); it finds a lower bound
1210 to the length of subject needed. It is not necessarily the greatest lower
1211 bound, but using it to avoid searching strings that are too short does give
1212 some useful speed-ups. The value is available to calling programs via
1213 pcre_fullinfo().
1214
1215 26. While implementing 25, I discovered to my embarrassment that pcretest had
1216 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1217 study optimizations had never been tested with that matching function.
1218 Oops. What is worse, even when it was passed study data, there was a bug in
1219 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1220 were also very few tests of studied patterns with pcre_dfa_exec().
1221
1222 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1223 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1224 on the other side of the coin, they are no longer allowed to have different
1225 names, because these cannot be distinguished in PCRE, and this has caused
1226 confusion. (This is a difference from Perl.)
1227
1228 28. When duplicate subpattern names are present (necessarily with different
1229 numbers, as required by 27 above), and a test is made by name in a
1230 conditional pattern, either for a subpattern having been matched, or for
1231 recursion in such a pattern, all the associated numbered subpatterns are
1232 tested, and the overall condition is true if the condition is true for any
1233 one of them. This is the way Perl works, and is also more like the way
1234 testing by number works.
1235
1236
1237 Version 7.9 11-Apr-09
1238 ---------------------
1239
1240 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1241 (pcretest), all targets were linked against these libraries. This included
1242 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1243 libraries. This caused unwanted dependencies to be created. This problem
1244 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1245 pcretest is linked with readline.
1246
1247 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1248 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1249 moved outside it again, because FALSE and TRUE are already defined in AIX,
1250 but BOOL is not.
1251
1252 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1253 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1254
1255 4. The pcregrep documentation said spaces were inserted as well as colons (or
1256 hyphens) following file names and line numbers when outputting matching
1257 lines. This is not true; no spaces are inserted. I have also clarified the
1258 wording for the --colour (or --color) option.
1259
1260 5. In pcregrep, when --colour was used with -o, the list of matching strings
1261 was not coloured; this is different to GNU grep, so I have changed it to be
1262 the same.
1263
1264 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1265 each matching line was coloured. Now it goes on to look for further matches
1266 of any of the test patterns, which is the same behaviour as GNU grep.
1267
1268 7. A pattern that could match an empty string could cause pcregrep to loop; it
1269 doesn't make sense to accept an empty string match in pcregrep, so I have
1270 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1271 seems to be how GNU grep behaves.
1272
1273 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1274 start or after a newline", because the conditional assertion was not being
1275 correctly handled. The rule now is that both the assertion and what follows
1276 in the first alternative must satisfy the test.
1277
1278 9. If auto-callout was enabled in a pattern with a conditional group whose
1279 condition was an assertion, PCRE could crash during matching, both with
1280 pcre_exec() and pcre_dfa_exec().
1281
1282 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1283 used for matching.
1284
1285 11. Unicode property support in character classes was not working for
1286 characters (bytes) greater than 127 when not in UTF-8 mode.
1287
1288 12. Added the -M command line option to pcretest.
1289
1290 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1291
1292 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1293
1294 16. Added comments and documentation about mis-use of no_arg in the C++
1295 wrapper.
1296
1297 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1298 from Martin Jerabek that uses macro names for all relevant character and
1299 string constants.
1300
1301 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1302 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1303 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1304 these, but not everybody uses configure.
1305
1306 19. A conditional group that had only one branch was not being correctly
1307 recognized as an item that could match an empty string. This meant that an
1308 enclosing group might also not be so recognized, causing infinite looping
1309 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1310 with the subject "ab", where knowledge that the repeated group can match
1311 nothing is needed in order to break the loop.
1312
1313 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1314 exec(), but without supplying a callout function, matching went wrong.
1315
1316 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1317 leak if the size of the offset vector was greater than 30. When the vector
1318 is smaller, the saved offsets during recursion go onto a local stack
1319 vector, but for larger vectors malloc() is used. It was failing to free
1320 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1321 error, in fact).
1322
1323 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1324 heapframe that is used only when UTF-8 support is enabled. This caused no
1325 problem, but was untidy.
1326
1327 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1328 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1329 included within another project.
1330
1331 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1332 slightly modified by me:
1333
1334 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1335 not building pcregrep.
1336
1337 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1338 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1339
1340 25. Forward references, both numeric and by name, in patterns that made use of
1341 duplicate group numbers, could behave incorrectly or give incorrect errors,
1342 because when scanning forward to find the reference group, PCRE was not
1343 taking into account the duplicate group numbers. A pattern such as
1344 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1345
1346 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1347 the feature of a custom pointer more persuasive (as requested by a user).
1348
1349 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1350 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1351 pre-defined.
1352
1353 28. Added support for (*UTF8) at the start of a pattern.
1354
1355 29. Arrange for flags added by the "release type" setting in CMake to be shown
1356 in the configuration summary.
1357
1358
1359 Version 7.8 05-Sep-08
1360 ---------------------
1361
1362 1. Replaced UCP searching code with optimized version as implemented for Ad
1363 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1364 stage table and inline lookup instead of a function, giving speed ups of 2
1365 to 5 times on some simple patterns that I tested. Permission was given to
1366 distribute the MultiStage2.py script that generates the tables (it's not in
1367 the tarball, but is in the Subversion repository).
1368
1369 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1370 scripts.
1371
1372 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1373 a group with a zero qualifier. The result of the study could be incorrect,
1374 or the function might crash, depending on the pattern.
1375
1376 4. Caseless matching was not working for non-ASCII characters in back
1377 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1378 It now works when Unicode Property Support is available.
1379
1380 5. In pcretest, an escape such as \x{de} in the data was always generating
1381 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1382 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1383 truncation.
1384
1385 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1386
1387 7. Added two (int) casts to pcregrep when printing the difference of two
1388 pointers, in case they are 64-bit values.
1389
1390 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1391 test 2 if it fails.
1392
1393 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1394 and a #define of that name to empty if it is not externally set. This is to
1395 allow users of MSVC to set it if necessary.
1396
1397 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1398 the convenience functions in the pcre_get.c source file.
1399
1400 11. An option change at the start of a pattern that had top-level alternatives
1401 could cause overwriting and/or a crash. This command provoked a crash in
1402 some environments:
1403
1404 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1405
1406 This potential security problem was recorded as CVE-2008-2371.
1407
1408 12. For a pattern where the match had to start at the beginning or immediately
1409 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1410 pcre_dfa_exec() could read past the end of the passed subject if there was
1411 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1412 pcretest so that it places the subject at the end of its malloc-ed buffer.
1413
1414 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1415 exec() might read past the end of the data buffer in UTF-8 mode.
1416
1417 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1418 the data contained the byte 0x85 as part of a UTF-8 character within its
1419 first line. This applied both to normal and DFA matching.
1420
1421 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1422 /^[^d]*?$/8 failed to match "abc".
1423
1424 16. Added a missing copyright notice to pcrecpp_internal.h.
1425
1426 17. Make it more clear in the documentation that values returned from
1427 pcre_exec() in ovector are byte offsets, not character counts.
1428
1429 18. Tidied a few places to stop certain compilers from issuing warnings.
1430
1431 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1432 supplied by Stefan Weber. I made a further small update for 7.8 because
1433 there is a change of source arrangements: the pcre_searchfuncs.c module is
1434 replaced by pcre_ucd.c.
1435
1436
1437 Version 7.7 07-May-08
1438 ---------------------
1439
1440 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1441 a string to a long long, pretend we don't even have a long long." This is
1442 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1443
1444 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1445 pre-7.6 versions, which defined a global no_arg variable instead of putting
1446 it in the RE class. (See also #8 below.)
1447
1448 3. Remove a line of dead code, identified by coverity and reported by Nuno
1449 Lopes.
1450
1451 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1452
1453 (1) The include/exclude patterns were being applied to the whole pathnames
1454 of files, instead of just to the final components.
1455
1456 (2) If there was more than one level of directory, the subdirectories were
1457 skipped unless they satisfied the include/exclude conditions. This is
1458 inconsistent with GNU grep (and could even be seen as contrary to the
1459 pcregrep specification - which I improved to make it absolutely clear).
1460 The action now is always to scan all levels of directory, and just
1461 apply the include/exclude patterns to regular files.
1462
1463 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1464 --exclude_dir in the tests to avoid scanning .svn directories.
1465
1466 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1467 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1468 doesn't support NULs in patterns.
1469
1470 7. Added some missing "const"s to declarations of static tables in
1471 pcre_compile.c and pcre_dfa_exec.c.
1472
1473 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1474 caused by fix #2 above. (Subsequently also a second patch to fix the
1475 first patch. And a third patch - this was a messy problem.)
1476
1477 9. Applied Craig's patch to remove the use of push_back().
1478
1479 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1480 matching function regexec().
1481
1482 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1483 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1484 references. PCRE supports relative numbers with this syntax (I don't think
1485 Oniguruma does).
1486
1487 12. Previously, a group with a zero repeat such as (...){0} was completely
1488 omitted from the compiled regex. However, this means that if the group
1489 was called as a subroutine from elsewhere in the pattern, things went wrong
1490 (an internal error was given). Such groups are now left in the compiled
1491 pattern, with a new opcode that causes them to be skipped at execution
1492 time.
1493
1494 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1495 to the way PCRE behaves:
1496
1497 (a) A lone ] character is dis-allowed (Perl treats it as data).
1498
1499 (b) A back reference to an unmatched subpattern matches an empty string
1500 (Perl fails the current match path).
1501
1502 (c) A data ] in a character class must be notated as \] because if the
1503 first data character in a class is ], it defines an empty class. (In
1504 Perl it is not possible to have an empty class.) The empty class []
1505 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1506 The negative empty class [^] matches any one character, independently
1507 of the DOTALL setting.
1508
1509 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1510 non-existent subpattern following a character class starting with ']' and
1511 containing () gave an internal compiling error instead of "reference to
1512 non-existent subpattern". Fortunately, when the pattern did exist, the
1513 compiled code was correct. (When scanning forwards to check for the
1514 existencd of the subpattern, it was treating the data ']' as terminating
1515 the class, so got the count wrong. When actually compiling, the reference
1516 was subsequently set up correctly.)
1517
1518 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1519 it was being rejected as not supported by pcre_dfa_exec(), even though
1520 other assertions are supported. I have made pcre_dfa_exec() support
1521 (*FAIL).
1522
1523 16. The implementation of 13c above involved the invention of a new opcode,
1524 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1525 cannot be changed at match time, I realized I could make a small
1526 improvement to matching performance by compiling OP_ALLANY instead of
1527 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1528 on the OP_ANY path.
1529
1530 17. Compiling pcretest on Windows with readline support failed without the
1531 following two fixes: (1) Make the unistd.h include conditional on
1532 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1533
1534 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1535 ncurses library to be included for pcretest when ReadLine support is
1536 requested, but also to allow for it to be overridden. This patch came from
1537 Daniel Bergström.
1538
1539 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1540 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1541 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1542 spotting this.
1543
1544
1545 Version 7.6 28-Jan-08
1546 ---------------------
1547
1548 1. A character class containing a very large number of characters with
1549 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1550 overflow.
1551
1552 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1553 HAVE_LONG_LONG is not defined.
1554
1555 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1556 bring them up to date and include new features. This patch includes:
1557
1558 - Fixed PH's badly added libz and libbz2 support.
1559 - Fixed a problem with static linking.
1560 - Added pcredemo. [But later removed - see 7 below.]
1561 - Fixed dftables problem and added an option.
1562 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1563 HAVE_LONG_LONG.
1564 - Added readline support for pcretest.
1565 - Added an listing of the option settings after cmake has run.
1566
1567 4. A user submitted a patch to Makefile that makes it easy to create
1568 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1569 Makefile.am that cause it to include this special target, without
1570 affecting anything else. Note that the same mingw target plus all
1571 the other distribution libraries and programs are now supported
1572 when configuring with CMake (see 6 below) instead of with
1573 Configure/Make.
1574
1575 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1576 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1577 exported in the Windows port". It has not yet been confirmed that the patch
1578 solves the problem, but it does no harm.
1579
1580 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1581 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1582 with CMake, and also correct the comment about stack recursion.
1583
1584 7. Remove the automatic building of pcredemo from the ./configure system and
1585 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1586 of a program that users should build themselves after PCRE is installed, so
1587 building it automatically is not really right. What is more, it gave
1588 trouble in some build environments.
1589
1590 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1591
1592
1593 Version 7.5 10-Jan-08
1594 ---------------------
1595
1596 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1597 values in parens when parsing an RE using the C++ wrapper."
1598
1599 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1600 Characters greater than 255 were excluded from the class instead of being
1601 included.
1602
1603 3. The same bug as (2) above applied to negated POSIX classes such as
1604 [:^space:].
1605
1606 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1607 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1608 I have changed it.
1609
1610 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1611 first named subpattern) and a construct such as (?&a) would reference the
1612 first named subpattern whose name started with "a" (in other words, the
1613 length check was missing). Both these problems are fixed. "Subpattern name
1614 expected" is now given for (?&) (a zero-length name), and this patch also
1615 makes it give the same error for \k'' (previously it complained that that
1616 was a reference to a non-existent subpattern).
1617
1618 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1619 this is right because (?- can be followed by option settings as well as by
1620 digits. I have, however, made the messages clearer.
1621
1622 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1623 than the number used in the conditional) now cause a compile-time error.
1624 This is actually not compatible with Perl, which accepts such patterns, but
1625 treats the conditional as always being FALSE (as PCRE used to), but it
1626 seems to me that giving a diagnostic is better.
1627
1628 8. Change "alphameric" to the more common word "alphanumeric" in comments
1629 and messages.
1630
1631 9. Fix two occurrences of "backslash" in comments that should have been
1632 "backspace".
1633
1634 10. Remove two redundant lines of code that can never be obeyed (their function
1635 was moved elsewhere).
1636
1637 11. The program that makes PCRE's Unicode character property table had a bug
1638 which caused it to generate incorrect table entries for sequences of
1639 characters that have the same character type, but are in different scripts.
1640 It amalgamated them into a single range, with the script of the first of
1641 them. In other words, some characters were in the wrong script. There were
1642 thirteen such cases, affecting characters in the following ranges:
1643
1644 U+002b0 - U+002c1
1645 U+0060c - U+0060d
1646 U+0061e - U+00612
1647 U+0064b - U+0065e
1648 U+0074d - U+0076d
1649 U+01800 - U+01805
1650 U+01d00 - U+01d77
1651 U+01d9b - U+01dbf
1652 U+0200b - U+0200f
1653 U+030fc - U+030fe
1654 U+03260 - U+0327f
1655 U+0fb46 - U+0fbb1
1656 U+10450 - U+1049d
1657
1658 12. The -o option (show only the matching part of a line) for pcregrep was not
1659 compatible with GNU grep in that, if there was more than one match in a
1660 line, it showed only the first of them. It now behaves in the same way as
1661 GNU grep.
1662
1663 13. If the -o and -v options were combined for pcregrep, it printed a blank
1664 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1665 does the same. The return code can be used to tell if there were any
1666 non-matching lines.
1667
1668 14. Added --file-offsets and --line-offsets to pcregrep.
1669
1670 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1671 infinitely looping recursion. The bug was that positive lookaheads were not
1672 being skipped when checking for a possible empty match (negative lookaheads
1673 and both kinds of lookbehind were skipped).
1674
1675 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1676 inclusion of <windows.h> to before rather than after the definition of
1677 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1678
1679 17. Specifying a possessive quantifier with a specific limit for a Unicode
1680 character property caused pcre_compile() to compile bad code, which led at
1681 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1682 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1683 caused the error; without that there was no problem.
1684
1685 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1686
1687 19. Added --enable-pcretest-libreadline.
1688
1689 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1690 RE::GlobalReplace(). As a result, the number of replacements returned was
1691 double what it should be. I removed one of the increments, but Craig sent a
1692 later patch that removed the other one (the right fix) and added unit tests
1693 that check the return values (which was not done before).
1694
1695 21. Several CMake things:
1696
1697 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1698 the names libpcre and libpcreposix, not just pcre and pcreposix.
1699
1700 (2) The above change means that pcretest and pcregrep are now correctly
1701 linked with the newly-built libraries, not previously installed ones.
1702
1703 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1704
1705 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1706 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1707 UTF-8 newline character). The key issue is that the pattern starts .*;
1708 this means that the match must be either at the beginning, or after a
1709 newline. The bug was in the code for advancing after a failed match and
1710 checking that the new position followed a newline. It was not taking
1711 account of UTF-8 characters correctly.
1712
1713 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1714 character classes. PCRE was not treating the sequence [:...:] as a
1715 character class unless the ... were all letters. Perl, however, seems to
1716 allow any characters between [: and :], though of course it rejects as
1717 unknown any "names" that contain non-letters, because all the known class
1718 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1719 for example, whereas PCRE did not - it did not recognize a POSIX character
1720 class. This seemed a bit dangerous, so the code has been changed to be
1721 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1722 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1723 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1724 Perl does, and where it didn't before.
1725
1726 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1727 Windows environments %n is disabled by default.
1728
1729
1730 Version 7.4 21-Sep-07
1731 ---------------------
1732
1733 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1734 means that a class such as [\s] counted as "explicit reference to CR or
1735 LF". That isn't really right - the whole point of the change was to try to
1736 help when there was an actual mention of one of the two characters. So now
1737 the change happens only if \r or \n (or a literal CR or LF) character is
1738 encountered.
1739
1740 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1741 of both had grown to the point where there were only 3 bits left.
1742 Fortunately, there was spare space in the data structure, and so I have
1743 moved the internal flags into a new 16-bit field to free up more option
1744 bits.
1745
1746 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1747 but did not set the internal JCHANGED flag - either of these is enough to
1748 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1749 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1750 start sets both bits.
1751
1752 4. Added options (at build time, compile time, exec time) to change \R from
1753 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1754
1755 5. doc/pcresyntax.html was missing from the distribution.
1756
1757 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1758 compatibility, even though it is no longer used.
1759
1760 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1761 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1762 windows.h file is present (where different names are used). [This was
1763 reversed later after testing - see 16 below.]
1764
1765 8. Changed all #include <config.h> to #include "config.h". There were also
1766 some further <pcre.h> cases that I changed to "pcre.h".
1767
1768 9. When pcregrep was used with the --colour option, it missed the line ending
1769 sequence off the lines that it output.
1770
1771 10. It was pointed out to me that arrays of string pointers cause lots of
1772 relocations when a shared library is dynamically loaded. A technique of
1773 using a single long string with a table of offsets can drastically reduce
1774 these. I have refactored PCRE in four places to do this. The result is
1775 dramatic:
1776
1777 Originally: 290
1778 After changing UCP table: 187
1779 After changing error message table: 43
1780 After changing table of "verbs" 36
1781 After changing table of Posix names 22
1782
1783 Thanks to the folks working on Gregex for glib for this insight.
1784
1785 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1786 unicode-properties was also set.
1787
1788 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1789
1790 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1791 checked only for CRLF.
1792
1793 14. Added casts to pcretest.c to avoid compiler warnings.
1794
1795 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1796
1797 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1798 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1799 entirely. This removes changes made in 7 above.
1800
1801 17. The CMake files have been updated, and there is now more information about
1802 building with CMake in the NON-UNIX-USE document.
1803
1804
1805 Version 7.3 28-Aug-07
1806 ---------------------
1807
1808 1. In the rejigging of the build system that eventually resulted in 7.1, the
1809 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1810 brackets there is not right, since it causes compilers to look for an
1811 installed pcre.h, not the version that is in the source that is being
1812 compiled (which of course may be different). I have changed it back to:
1813
1814 #include "pcre.h"
1815
1816 I have a vague recollection that the change was concerned with compiling in
1817 different directories, but in the new build system, that is taken care of
1818 by the VPATH setting the Makefile.
1819
1820 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1821 when the subject happened to end in the byte 0x85 (e.g. if the last
1822 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1823 characters but of course it shouldn't be taken as a newline when it is part
1824 of another character. The bug was that, for an unlimited repeat of . in
1825 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1826 characters when looking for a newline.
1827
1828 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1829
1830 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1831 in debug output.
1832
1833 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1834 long printing in the pcrecpp unittest when running under MinGW.
1835
1836 6. ESC_K was left out of the EBCDIC table.
1837
1838 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1839 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1840 limit also applies to "virtual nesting" when a pattern is recursive, and in
1841 this case 1000 isn't so big. I have been able to remove this limit at the
1842 expense of backing off one optimization in certain circumstances. Normally,
1843 when pcre_exec() would call its internal match() function recursively and
1844 immediately return the result unconditionally, it uses a "tail recursion"
1845 feature to save stack. However, when a subpattern that can match an empty
1846 string has an unlimited repetition quantifier, it no longer makes this
1847 optimization. That gives it a stack frame in which to save the data for
1848 checking that an empty string has been matched. Previously this was taken
1849 from the 1000-entry workspace that had been reserved. So now there is no
1850 explicit limit, but more stack is used.
1851
1852 8. Applied Daniel's patches to solve problems with the import/export magic
1853 syntax that is required for Windows, and which was going wrong for the
1854 pcreposix and pcrecpp parts of the library. These were overlooked when this
1855 problem was solved for the main library.
1856
1857 9. There were some crude static tests to avoid integer overflow when computing
1858 the size of patterns that contain repeated groups with explicit upper
1859 limits. As the maximum quantifier is 65535, the maximum group length was
1860 set at 30,000 so that the product of these two numbers did not overflow a
1861 32-bit integer. However, it turns out that people want to use groups that
1862 are longer than 30,000 bytes (though not repeat them that many times).
1863 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1864 made it possible to implement the integer overflow checks in a much more
1865 dynamic way, which I have now done. The artificial limitation on group
1866 length has been removed - we now have only the limit on the total length of
1867 the compiled pattern, which depends on the LINK_SIZE setting.
1868
1869 10. Fixed a bug in the documentation for get/copy named substring when
1870 duplicate names are permitted. If none of the named substrings are set, the
1871 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1872 empty string.
1873
1874 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1875 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1876 because the ] is interpreted as the first data character and the
1877 terminating ] is not found. PCRE has been made compatible with Perl in this
1878 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1879 cause memory overwriting.
1880
1881 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1882 string has been matched (to stop an infinite loop). It was not recognizing
1883 a conditional subpattern that could match an empty string if that
1884 subpattern was within another subpattern. For example, it looped when
1885 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1886 condition was not nested. This bug has been fixed.
1887
1888 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1889 past the start of the subject in the presence of bytes with the top bit
1890 set, for example "\x8aBCD".
1891
1892 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1893 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1894
1895 14. Optimized (?!) to (*FAIL).
1896
1897 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1898 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1899 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1900 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1901 does: it's just the validity check that is more restrictive.
1902
1903 16. Inserted checks for integer overflows during escape sequence (backslash)
1904 processing, and also fixed erroneous offset values for syntax errors during
1905 backslash processing.
1906
1907 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1908 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1909
1910 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1911 caused an overrun.
1912
1913 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1914 something other than just ASCII characters) inside a group that had an
1915 unlimited repeat caused a loop at compile time (while checking to see
1916 whether the group could match an empty string).
1917
1918 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1919 [\P{Any}] did so. (Error in the code for printing property names.)
1920
1921 21. An orphan \E inside a character class could cause a crash.
1922
1923 22. A repeated capturing bracket such as (A)? could cause a wild memory
1924 reference during compilation.
1925
1926 23. There are several functions in pcre_compile() that scan along a compiled
1927 expression for various reasons (e.g. to see if it's fixed length for look
1928 behind). There were bugs in these functions when a repeated \p or \P was
1929 present in the pattern. These operators have additional parameters compared
1930 with \d, etc, and these were not being taken into account when moving along
1931 the compiled data. Specifically:
1932
1933 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1934 length.
1935
1936 (b) An item such as \pL+ within a repeated group could cause crashes or
1937 loops.
1938
1939 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1940 "reference to non-existent subpattern" error.
1941
1942 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1943
1944 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1945 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1946
1947 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1948
1949 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1950 character were causing crashes (broken optimization).
1951
1952 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1953 \p or \P) caused a compile-time loop.
1954
1955 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1956 break. For example, the unstudied pattern [\r\n]A does not match the string
1957 "\r\nA" because change 7.0/46 below moves the current point on by two
1958 characters after failing to match at the start. However, the pattern \nA
1959 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1960 the same is true. There doesn't seem any very clean way out of this, but
1961 what I have chosen to do makes the common cases work: PCRE now takes note
1962 of whether there can be an explicit match for \r or \n anywhere in the
1963 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1964 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1965 pattern has explicit CR or LF references.
1966
1967 29. Added (*CR) etc for changing newline setting at start of pattern.
1968
1969
1970 Version 7.2 19-Jun-07
1971 ---------------------
1972
1973 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1974 which is apparently normally available under Windows.
1975
1976 2. Re-jig the pcregrep tests with different newline settings in an attempt
1977 to make them independent of the local environment's newline setting.
1978
1979 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1980
1981 4. Some of the "internals" tests were previously cut out when the link size
1982 was not 2, because the output contained actual offsets. The recent new
1983 "Z" feature of pcretest means that these can be cut out, making the tests
1984 usable with all link sizes.
1985
1986 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1987 stack recursion. This gives a massive performance boost under BSD, but just
1988 a small improvement under Linux. However, it saves one field in the frame
1989 in all cases.
1990
1991 6. Added more features from the forthcoming Perl 5.10:
1992
1993 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1994 recursion call. It refers to the nth most recently opened parentheses.
1995
1996 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1997 to be opened parentheses.
1998
1999 (c) Conditions that refer to capturing parentheses can be specified
2000 relatively, for example, (?(-2)... or (?(+3)...
2001
2002 (d) \K resets the start of the current match so that everything before
2003 is not part of it.
2004
2005 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2006
2007 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2008 reference syntax.
2009
2010 (g) (?| introduces a group in which the numbering of parentheses in each
2011 alternative starts with the same number.
2012
2013 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2014
2015 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2016 PCRE_INFO_JCHANGED.
2017
2018 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2019 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2020 for detecting groups that can match an empty string.
2021
2022 9. A pattern with a very large number of alternatives (more than several
2023 hundred) was running out of internal workspace during the pre-compile
2024 phase, where pcre_compile() figures out how much memory will be needed. A
2025 bit of new cunning has reduced the workspace needed for groups with
2026 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2027 workspace instead of running out of the 4096 that are available.
2028
2029 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2030
2031 11. Applied patch from Google to remove an optimization that didn't quite work.
2032 The report of the bug said:
2033
2034 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2035 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2036 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2037
2038 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2039 it matched the wrong number of bytes.
2040
2041
2042 Version 7.1 24-Apr-07
2043 ---------------------
2044
2045 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2046 that is more "standard", making use of automake and other Autotools. There
2047 is some re-arrangement of the files and adjustment of comments consequent
2048 on this.
2049
2050 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2051 for recursive directory scanning broke on some systems because the files
2052 are not scanned in any specific order and on different systems the order
2053 was different. A call to "sort" has been inserted into RunGrepTest for the
2054 approprate test as a short-term fix. In the longer term there may be an
2055 alternative.
2056
2057 3. I had an email from Eric Raymond about problems translating some of PCRE's
2058 man pages to HTML (despite the fact that I distribute HTML pages, some
2059 people do their own conversions for various reasons). The problems
2060 concerned the use of low-level troff macros .br and .in. I have therefore
2061 removed all such uses from the man pages (some were redundant, some could
2062 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2063 HTML has been updated to handle .nf/.fi and to complain if it encounters
2064 .br or .in.
2065
2066 4. Updated comments in configure.ac that get placed in config.h.in and also
2067 arranged for config.h to be included in the distribution, with the name
2068 config.h.generic, for the benefit of those who have to compile without
2069 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2070
2071 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2072 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2073 makevp.bat for the current PCRE, using the additional files
2074 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2075
2076 6. A Windows user reported a minor discrepancy with test 2, which turned out
2077 to be caused by a trailing space on an input line that had got lost in his
2078 copy. The trailing space was an accident, so I've just removed it.
2079
2080 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2081 that is needed.
2082
2083 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2084 as "const" (a) because they are and (b) because it helps the PHP
2085 maintainers who have recently made a script to detect big data structures
2086 in the php code that should be moved to the .rodata section. I remembered
2087 to update Builducptable as well, so it won't revert if ucptable.h is ever
2088 re-created.
2089
2090 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2091 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2092 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2093 support is not required. This saves 1.5-2K of code, which is important in
2094 some applications.
2095
2096 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2097 so as not to refer to the tables, even though these functions will never be
2098 called when UTF-8 support is disabled. Otherwise there are problems with a
2099 shared library.
2100
2101 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2102
2103 (a) It was defining its arguments as char * instead of void *.
2104
2105 (b) It was assuming that all moves were upwards in memory; this was true
2106 a long time ago when I wrote it, but is no longer the case.
2107
2108 The emulated memove() is provided for those environments that have neither
2109 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2110 is clearly not the case, as these two bugs were recently reported.
2111
2112 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2113 and Detrail to create the HTML documentation, the .txt form of the man
2114 pages, and it removes trailing spaces from listed files. It also creates
2115 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2116 case, it wraps all the #defines with #ifndefs. This script should be run
2117 before "make dist".
2118
2119 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2120 with Unicode property support.
2121
2122 (a) For a maximizing quantifier, if the two different cases of the
2123 character were of different lengths in their UTF-8 codings (there are
2124 some cases like this - I found 11), and the matching function had to
2125 back up over a mixture of the two cases, it incorrectly assumed they
2126 were both the same length.
2127
2128 (b) When PCRE was configured to use the heap rather than the stack for
2129 recursion during matching, it was not correctly preserving the data for
2130 the other case of a UTF-8 character when checking ahead for a match
2131 while processing a minimizing repeat. If the check also involved
2132 matching a wide character, but failed, corruption could cause an
2133 erroneous result when trying to check for a repeat of the original
2134 character.
2135
2136 13. Some tidying changes to the testing mechanism:
2137
2138 (a) The RunTest script now detects the internal link size and whether there
2139 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2140 values substituted by "configure". (The RunGrepTest script already did
2141 this for UTF-8.) The configure.ac script no longer substitutes the
2142 relevant variables.
2143
2144 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2145 with length and offset values. This means that the output is different
2146 for different internal link sizes. Test 2 is skipped for link sizes
2147 other than 2 because of this, bypassing the problem. Unfortunately,
2148 there was also a test in test 3 (the locale tests) that used /B and
2149 failed for link sizes other than 2. Rather than cut the whole test out,
2150 I have added a new /Z option to pcretest that replaces the length and
2151 offset values with spaces. This is now used to make test 3 independent
2152 of link size. (Test 2 will be tidied up later.)
2153
2154 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2155 segmentation fault instead of returning the appropriate error message.
2156
2157 15. In multiline mode when the newline sequence was set to "any", the pattern
2158 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2159 This doesn't seem right; it now treats the CRLF combination as the line
2160 ending, and so does not match in that case. It's only a pattern such as ^$
2161 that would hit this one: something like ^ABC$ would have failed after \r
2162 and then tried again after \r\n.
2163
2164 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2165 in an attempt to make files that differ only in their line terminators
2166 compare equal. This works on Linux.
2167
2168 17. Under certain error circumstances pcregrep might try to free random memory
2169 as it exited. This is now fixed, thanks to valgrind.
2170
2171 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2172 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2173 was because its rules for how to advance for /g after matching an empty
2174 string at the end of a line did not allow for this case. They now check for
2175 it specially.
2176
2177 20. pcretest is supposed to handle patterns and data of any length, by
2178 extending its buffers when necessary. It was getting this wrong when the
2179 buffer for a data line had to be extended.
2180
2181 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2182 CRLF as a newline sequence.
2183
2184 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2185 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2186 I have nevertheless tidied it up.
2187
2188 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2189
2190 24. Added a man page for pcre-config.
2191
2192
2193 Version 7.0 19-Dec-06
2194 ---------------------
2195
2196 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2197 moving to gcc 4.1.1.
2198
2199 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2200 sys/time.h, which is documented as needed for this function. It doesn't
2201 seem to matter on Linux, but it showed up on some releases of OS X.
2202
2203 3. It seems that there are systems where bytes whose values are greater than
2204 127 match isprint() in the "C" locale. The "C" locale should be the
2205 default when a C program starts up. In most systems, only ASCII printing
2206 characters match isprint(). This difference caused the output from pcretest
2207 to vary, making some of the tests fail. I have changed pcretest so that:
2208
2209 (a) When it is outputting text in the compiled version of a pattern, bytes
2210 other than 32-126 are always shown as hex escapes.
2211
2212 (b) When it is outputting text that is a matched part of a subject string,
2213 it does the same, unless a different locale has been set for the match
2214 (using the /L modifier). In this case, it uses isprint() to decide.
2215
2216 4. Fixed a major bug that caused incorrect computation of the amount of memory
2217 required for a compiled pattern when options that changed within the
2218 pattern affected the logic of the preliminary scan that determines the
2219 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2220 that the computed length was too small. The symptoms of this bug were
2221 either the PCRE error "internal error: code overflow" from pcre_compile(),
2222 or a glibc crash with a message such as "pcretest: free(): invalid next
2223 size (fast)". Examples of patterns that provoked this bug (shown in
2224 pcretest format) are:
2225
2226 /(?-x: )/x
2227 /(?x)(?-x: \s*#\s*)/
2228 /((?i)[\x{c0}])/8
2229 /(?i:[\x{c0}])/8
2230
2231 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2232 is now done differently.
2233
2234 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2235 wrapper classes; (b) implement a new function in the C++ scanner that is
2236 more efficient than the old way of doing things because it avoids levels of
2237 recursion in the regex matching; (c) add a paragraph to the documentation
2238 for the FullMatch() function.
2239
2240 6. The escape sequence \n was being treated as whatever was defined as
2241 "newline". Not only was this contrary to the documentation, which states
2242 that \n is character 10 (hex 0A), but it also went horribly wrong when
2243 "newline" was defined as CRLF. This has been fixed.
2244
2245 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2246 was being set to -1 for the "end of line" case (supposedly a value that no
2247 character can have). Though this value is never used (the check for end of
2248 line is "zero bytes in current character"), it caused compiler complaints.
2249 I've changed it to 0xffffffff.
2250
2251 8. In pcre_version.c, the version string was being built by a sequence of
2252 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2253 string (as it is for production releases) called a macro with an empty
2254 argument. The C standard says the result of this is undefined. The gcc
2255 compiler treats it as an empty string (which was what was wanted) but it is
2256 reported that Visual C gives an error. The source has been hacked around to
2257 avoid this problem.
2258
2259 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2260 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2261 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2262 of them did).
2263
2264 10. Originally, pcretest opened its input and output without "b"; then I was
2265 told that "b" was needed in some environments, so it was added for release
2266 5.0 to both the input and output. (It makes no difference on Unix-like
2267 systems.) Later I was told that it is wrong for the input on Windows. I've
2268 now abstracted the modes into two macros, to make it easier to fiddle with
2269 them, and removed "b" from the input mode under Windows.
2270
2271 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2272
2273 12. Added -help and --help to pcretest as an official way of being reminded
2274 of the options.
2275
2276 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2277 and pcrecpp.cc because they annoy compilers at high warning levels.
2278
2279 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2280
2281 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2282 scripts are not C programs :-) and which was not noticed because it works
2283 on Linux.
2284
2285 16. pcretest is supposed to handle any length of pattern and data line (as one
2286 line or as a continued sequence of lines) by extending its input buffer if
2287 necessary. This feature was broken for very long pattern lines, leading to
2288 a string of junk being passed to pcre_compile() if the pattern was longer
2289 than about 50K.
2290
2291 17. I have done a major re-factoring of the way pcre_compile() computes the
2292 amount of memory needed for a compiled pattern. Previously, there was code
2293 that made a preliminary scan of the pattern in order to do this. That was
2294 OK when PCRE was new, but as the facilities have expanded, it has become
2295 harder and harder to keep it in step with the real compile phase, and there
2296 have been a number of bugs (see for example, 4 above). I have now found a
2297 cunning way of running the real compile function in a "fake" mode that
2298 enables it to compute how much memory it would need, while actually only
2299 ever using a few hundred bytes of working memory and without too many
2300 tests of the mode. This should make future maintenance and development
2301 easier. A side effect of this work is that the limit of 200 on the nesting
2302 depth of parentheses has been removed (though this was never a serious
2303 limitation, I suspect). However, there is a downside: pcre_compile() now
2304 runs more slowly than before (30% or more, depending on the pattern). I
2305 hope this isn't a big issue. There is no effect on runtime performance.
2306
2307 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2308 newline (only possible for the last line of a file) and it was a
2309 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2310
2311 19. Added additional timing features to pcretest. (1) The -tm option now times
2312 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2313 separate command line item, by a number that specifies the number of
2314 repeats to use when timing. The default is 50000; this gives better
2315 precision, but takes uncomfortably long for very large patterns.
2316
2317 20. Extended pcre_study() to be more clever in cases where a branch of a
2318 subpattern has no definite first character. For example, (a*|b*)[cd] would
2319 previously give no result from pcre_study(). Now it recognizes that the
2320 first character must be a, b, c, or d.
2321
2322 21. There was an incorrect error "recursive call could loop indefinitely" if
2323 a subpattern (or the entire pattern) that was being tested for matching an
2324 empty string contained only one non-empty item after a nested subpattern.
2325 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2326 incorrectly, because the \d was being skipped in the check.
2327
2328 22. The pcretest program now has a new pattern option /B and a command line
2329 option -b, which is equivalent to adding /B to every pattern. This causes
2330 it to show the compiled bytecode, without the additional information that
2331 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2332 is the same as /B/I).
2333
2334 23. A new optimization is now able automatically to treat some sequences such
2335 as a*b as a*+b. More specifically, if something simple (such as a character
2336 or a simple class like \d) has an unlimited quantifier, and is followed by
2337 something that cannot possibly match the quantified thing, the quantifier
2338 is automatically "possessified".
2339
2340 24. A recursive reference to a subpattern whose number was greater than 39
2341 went wrong under certain circumstances in UTF-8 mode. This bug could also
2342 have affected the operation of pcre_study().
2343
2344 25. Realized that a little bit of performance could be had by replacing
2345 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2346
2347 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2348
2349 27. Possessive quantifiers such as a++ were previously implemented by turning
2350 them into atomic groups such as ($>a+). Now they have their own opcodes,
2351 which improves performance. This includes the automatically created ones
2352 from 23 above.
2353
2354 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2355 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2356 the first matched character to be a colon. This applied both to named and
2357 numbered groups.
2358
2359 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2360
2361 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2362 building PCRE on Windows easier, so I have included it in the distribution.
2363
2364 31. There is now a check in pcretest against a ridiculously large number being
2365 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2366 loop, the loop is abandoned.
2367
2368 32. Forward references to subpatterns in conditions such as (?(2)...) where
2369 subpattern 2 is defined later cause pcre_compile() to search forwards in
2370 the pattern for the relevant set of parentheses. This search went wrong
2371 when there were unescaped parentheses in a character class, parentheses
2372 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2373
2374 33. "Subroutine" calls and backreferences were previously restricted to
2375 referencing subpatterns earlier in the regex. This restriction has now
2376 been removed.
2377
2378 34. Added a number of extra features that are going to be in Perl 5.10. On the
2379 whole, these are just syntactic alternatives for features that PCRE had
2380 previously implemented using the Python syntax or my own invention. The
2381 other formats are all retained for compatibility.
2382
2383 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2384 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2385 also .NET compatible.
2386
2387 (b) A recursion or subroutine call to a named group can now be defined as
2388 (?&name) as well as (?P>name).
2389
2390 (c) A backreference to a named group can now be defined as \k<name> or
2391 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2392 5.10, are also .NET compatible.
2393
2394 (d) A conditional reference to a named group can now use the syntax
2395 (?(<name>) or (?('name') as well as (?(name).
2396
2397 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2398 groups (named and numbered) that are never evaluated inline, but can be
2399 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2400 is always false. There may be only one alternative in such a group.
2401
2402 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2403 as the simple (?(R). The condition is true only if the most recent
2404 recursion is that of the given number or name. It does not search out
2405 through the entire recursion stack.
2406
2407 (g) The escape \gN or \g{N} has been added, where N is a positive or
2408 negative number, specifying an absolute or relative reference.
2409
2410 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2411 some "unreachable code" warnings.
2412
2413 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2414 things, this adds five new scripts.
2415
2416 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2417 There were also incompatibilities regarding the handling of \Q..\E inside
2418 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2419 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2420
2421 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2422 matches an empty string, and forcibly breaks the loop. There were bugs in
2423 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2424 against aaaa the result was just "a" rather than "aaaa", for example. Two
2425 separate and independent bugs (that affected different cases) have been
2426 fixed.
2427
2428 39. Refactored the code to abolish the use of different opcodes for small
2429 capturing bracket numbers. This is a tidy that I avoided doing when I
2430 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2431 The new approach is not only tidier, it makes it possible to reduce the
2432 memory needed to fix the previous bug (38).
2433
2434 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2435 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2436 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2437 mode.
2438
2439 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2440 report.
2441
2442 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2443 copy construction and assignment in the C++ wrapper.
2444
2445 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2446 couple of bugs that could have given wrong results in the "--newline=crlf"
2447 case.
2448
2449 44. Added a number of casts and did some reorganization of signed/unsigned int
2450 variables following suggestions from Dair Grant. Also renamed the variable
2451 "this" as "item" because it is a C++ keyword.
2452
2453 45. Arranged for dftables to add
2454
2455 #include "pcre_internal.h"
2456
2457 to pcre_chartables.c because without it, gcc 4.x may remove the array
2458 definition from the final binary if PCRE is built into a static library and
2459 dead code stripping is activated.
2460
2461 46. For an unanchored pattern, if a match attempt fails at the start of a
2462 newline sequence, and the newline setting is CRLF or ANY, and the next two
2463 characters are CRLF, advance by two characters instead of one.
2464
2465
2466 Version 6.7 04-Jul-06
2467 ---------------------
2468
2469 1. In order to handle tests when input lines are enormously long, pcretest has
2470 been re-factored so that it automatically extends its buffers when
2471 necessary. The code is crude, but this _is_ just a test program. The
2472 default size has been increased from 32K to 50K.
2473
2474 2. The code in pcre_study() was using the value of the re argument before
2475 testing it for NULL. (Of course, in any sensible call of the function, it
2476 won't be NULL.)
2477
2478 3. The memmove() emulation function in pcre_internal.h, which is used on
2479 systems that lack both memmove() and bcopy() - that is, hardly ever -
2480 was missing a "static" storage class specifier.
2481
2482 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2483 containing an extended class (one that cannot be represented by a bitmap
2484 because it contains high-valued characters or Unicode property items, e.g.
2485 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2486 pattern, but PCRE should not loop if you do not (it no longer does).
2487 [Detail: two cases were found: (a) a repeated subpattern containing an
2488 extended class; (b) a recursive reference to a subpattern that followed a
2489 previous extended class. It wasn't skipping over the extended class
2490 correctly when UTF-8 mode was not set.]
2491
2492 5. A negated single-character class was not being recognized as fixed-length
2493 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2494 compile error "lookbehind assertion is not fixed length".
2495
2496 6. The RunPerlTest auxiliary script was showing an unexpected difference
2497 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2498 write a Perl script that can interpret lines of an input file either as
2499 byte characters or as UTF-8, which is what "perltest" was being required to
2500 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2501 can't do is switch easily at run time between having the "use utf8;" pragma
2502 or not. In the end, I fudged it by using the RunPerlTest script to insert
2503 "use utf8;" explicitly for the UTF-8 tests.
2504
2505 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2506 the end of the subject string, contrary to the documentation and to what
2507 Perl does. This was true of both matching functions. Now it matches only at
2508 the start of the subject and immediately after *internal* newlines.
2509
2510 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2511 a pointer to an int instead of a pointer to an unsigned long int. This
2512 caused problems on 64-bit systems.
2513
2514 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2515 instance of the 'standard' template library not being so standard".
2516
2517 10. There was no check on the number of named subpatterns nor the maximum
2518 length of a subpattern name. The product of these values is used to compute
2519 the size of the memory block for a compiled pattern. By supplying a very
2520 long subpattern name and a large number of named subpatterns, the size
2521 computation could be caused to overflow. This is now prevented by limiting
2522 the length of names to 32 characters, and the number of named subpatterns
2523 to 10,000.
2524
2525 11. Subpatterns that are repeated with specific counts have to be replicated in
2526 the compiled pattern. The size of memory for this was computed from the
2527 length of the subpattern and the repeat count. The latter is limited to
2528 65535, but there was no limit on the former, meaning that integer overflow
2529 could in principle occur. The compiled length of a repeated subpattern is
2530 now limited to 30,000 bytes in order to prevent this.
2531
2532 12. Added the optional facility to have named substrings with the same name.
2533
2534 13. Added the ability to use a named substring as a condition, using the
2535 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2536 are numbers (not recommended). Forward references are permitted.
2537
2538 14. Added forward references in named backreferences (if you see what I mean).
2539
2540 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2541 pattern could run off the end of the subject. For example, the pattern
2542 "(?s)(.{1,5})"8 did this with the subject "ab".
2543
2544 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2545 PCRE_CASELESS was set when matching characters that were quantified with ?
2546 or *.
2547
2548 17. A character class other than a single negated character that had a minimum
2549 but no maximum quantifier - for example [ab]{6,} - was not handled
2550 correctly by pce_dfa_exec(). It would match only one character.
2551
2552 18. A valid (though odd) pattern that looked like a POSIX character
2553 class but used an invalid character after [ (for example [[,abc,]]) caused
2554 pcre_compile() to give the error "Failed: internal error: code overflow" or
2555 in some cases to crash with a glibc free() error. This could even happen if
2556 the pattern terminated after [[ but there just happened to be a sequence of
2557 letters, a binary zero, and a closing ] in the memory that followed.
2558
2559 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2560 over the years. Originally (before any Unicode support), just the bottom 8
2561 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2562 output from "man perlunicode" includes this:
2563
2564 The regular expression compiler produces polymorphic opcodes. That
2565 is, the pattern adapts to the data and automatically switches to
2566 the Unicode character scheme when presented with Unicode data--or
2567 instead uses a traditional byte scheme when presented with byte
2568 data.
2569
2570 Sadly, a wide octal escape does not cause a switch, and in a string with
2571 no other multibyte characters, these octal escapes are treated as before.
2572 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2573 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2574 Unicode string.
2575
2576 I have not perpetrated such confusion in PCRE. Up till now, it took just
2577 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2578 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2579 translate to the appropriate multibyte character.
2580
2581 29. Applied some refactoring to reduce the number of warnings from Microsoft
2582 and Borland compilers. This has included removing the fudge introduced
2583 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2584 a warning about an unused variable.
2585
2586 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2587 characters since release 4.0, because Perl (from release 5.004) does not.
2588 [Or at least, is documented not to: some releases seem to be in conflict
2589 with the documentation.] However, when a pattern was studied with
2590 pcre_study() and all its branches started with \s, PCRE still included VT
2591 as a possible starting character. Of course, this did no harm; it just
2592 caused an unnecessary match attempt.
2593
2594 22. Removed a now-redundant internal flag bit that recorded the fact that case
2595 dependency changed within the pattern. This was once needed for "required
2596 byte" processing, but is no longer used. This recovers a now-scarce options
2597 bit. Also moved the least significant internal flag bit to the most-
2598 significant bit of the word, which was not previously used (hangover from
2599 the days when it was an int rather than a uint) to free up another bit for
2600 the future.
2601
2602 23. Added support for CRLF line endings as well as CR and LF. As well as the
2603 default being selectable at build time, it can now be changed at runtime
2604 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2605 specify that it is scanning data with non-default line endings.
2606
2607 24. Changed the definition of CXXLINK to make it agree with the definition of
2608 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2609
2610 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2611 recursions. This makes a big different to stack usage for some patterns.
2612
2613 26. If a subpattern containing a named recursion or subroutine reference such
2614 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2615 the space required for the compiled pattern went wrong and gave too small a
2616 value. Depending on the environment, this could lead to "Failed: internal
2617 error: code overflow at offset 49" or "glibc detected double free or
2618 corruption" errors.
2619
2620 27. Applied patches from Google (a) to support the new newline modes and (b) to
2621 advance over multibyte UTF-8 characters in GlobalReplace.
2622
2623 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2624 difference for some implementation of PCRE in some Windows version.
2625
2626 29. Added some extra testing facilities to pcretest:
2627
2628 \q<number> in a data line sets the "match limit" value
2629 \Q<number> in a data line sets the "match recursion limt" value
2630 -S <number> sets the stack size, where <number> is in megabytes
2631
2632 The -S option isn't available for Windows.
2633
2634
2635 Version 6.6 06-Feb-06
2636 ---------------------
2637
2638 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2639 in pcreposix.h. I have copied the definition from pcre.h.
2640
2641 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2642 because pcre.h is no longer a built file.
2643
2644 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2645 not normally included in the compiled code.
2646
2647
2648 Version 6.5 01-Feb-06
2649 ---------------------
2650
2651 1. When using the partial match feature with pcre_dfa_exec(), it was not
2652 anchoring the second and subsequent partial matches at the new starting
2653 point. This could lead to incorrect results. For example, with the pattern
2654 /1234/, partially matching against "123" and then "a4" gave a match.
2655
2656 2. Changes to pcregrep:
2657
2658 (a) All non-match returns from pcre_exec() were being treated as failures
2659 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2660 error message is output. Some extra information is given for the
2661 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2662 probably the only errors that are likely to be caused by users (by
2663 specifying a regex that has nested indefinite repeats, for instance).
2664 If there are more than 20 of these errors, pcregrep is abandoned.
2665
2666 (b) A binary zero was treated as data while matching, but terminated the
2667 output line if it was written out. This has been fixed: binary zeroes
2668 are now no different to any other data bytes.
2669
2670 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2671 used to set a locale for matching. The --locale=xxxx long option has
2672 been added (no short equivalent) to specify a locale explicitly on the
2673 pcregrep command, overriding the environment variables.
2674
2675 (d) When -B was used with -n, some line numbers in the output were one less
2676 than they should have been.
2677
2678 (e) Added the -o (--only-matching) option.
2679
2680 (f) If -A or -C was used with -c (count only), some lines of context were
2681 accidentally printed for the final match.
2682
2683 (g) Added the -H (--with-filename) option.
2684
2685 (h) The combination of options -rh failed to suppress file names for files
2686 that were found from directory arguments.
2687
2688 (i) Added the -D (--devices) and -d (--directories) options.
2689
2690 (j) Added the -F (--fixed-strings) option.
2691
2692 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2693
2694 (l) Added the --colo(u)r option.
2695
2696 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2697 is not present by default.
2698
2699 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2700 items such as (?R) or (?1), when the recursion could match a number of
2701 alternatives. If it matched one of the alternatives, but subsequently,
2702 outside the recursion, there was a failure, the code tried to back up into
2703 the recursion. However, because of the way PCRE is implemented, this is not
2704 possible, and the result was an incorrect result from the match.
2705
2706 In order to prevent this happening, the specification of recursion has
2707 been changed so that all such subpatterns are automatically treated as
2708 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2709
2710 4. I had overlooked the fact that, in some locales, there are characters for
2711 which isalpha() is true but neither isupper() nor islower() are true. In
2712 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2713 and ordfeminine) are like this. This affected the treatment of \w and \W
2714 when they appeared in character classes, but not when they appeared outside
2715 a character class. The bit map for "word" characters is now created
2716 separately from the results of isalnum() instead of just taking it from the
2717 upper, lower, and digit maps. (Plus the underscore character, of course.)
2718
2719 5. The above bug also affected the handling of POSIX character classes such as
2720 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2721 permanent tables. Instead, the bit maps for such a class were previously
2722 created as the appropriate unions of the upper, lower, and digit bitmaps.
2723 Now they are created by subtraction from the [[:word:]] class, which has
2724 its own bitmap.
2725
2726 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2727 It is created by subtracting the vertical space characters (\x09, \x0a,
2728 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2729 subtraction was done in the overall bitmap for a character class, meaning
2730 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2731 be recognized. This bug has been fixed.
2732
2733 7. Patches from the folks at Google:
2734
2735 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2736 real life, but is still worth protecting against".
2737
2738 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2739 regular expressions".
2740
2741 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2742 have it.
2743
2744 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2745 "configure" and the latter not, in order to fix a problem somebody had
2746 with compiling the Arg class on HP-UX.
2747
2748 (e) Improve the error-handling of the C++ wrapper a little bit.
2749
2750 (f) New tests for checking recursion limiting.
2751
2752 8. The pcre_memmove() function, which is used only if the environment does not
2753 have a standard memmove() function (and is therefore rarely compiled),
2754 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2755 returning a result (though PCRE never actually uses the result).
2756
2757 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2758 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2759 returned instead of calling malloc() with an overflowing number that would
2760 most likely cause subsequent chaos.
2761
2762 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2763
2764 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2765 with this option is matched, the nmatch and pmatch options of regexec() are
2766 ignored.
2767
2768 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2769 provided in case anyone wants to the the POSIX interface with UTF-8
2770 strings.
2771
2772 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2773 C++ linking (needed for some HP-UX environments).
2774
2775 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2776 (unused parameter) and in the pcre_printint() function (omitted "default"
2777 switch label when the default is to do nothing).
2778
2779 15. Added some code to make it possible, when PCRE is compiled as a C++
2780 library, to replace subject pointers for pcre_exec() with a smart pointer
2781 class, thus making it possible to process discontinuous strings.
2782
2783 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2784 much the same function. They were added by different people who were trying
2785 to make PCRE easy to compile on non-Unix systems. It has been suggested
2786 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2787 for compiling on Windows systems. I have therefore replaced it with
2788 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2789 defaults to "extern" for C or "extern C" for C++, which works fine on
2790 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2791 SCOPE with something explicit in config.h. In addition:
2792
2793 (a) pcreposix.h still had just "extern" instead of either of these macros;
2794 I have replaced it with PCRE_DATA_SCOPE.
2795
2796 (b) Functions such as _pcre_xclass(), which are internal to the library,
2797 but external in the C sense, all had PCRE_EXPORT in their definitions.
2798 This is apparently wrong for the Windows case, so I have removed it.
2799 (It makes no difference on Unix-like systems.)
2800
2801 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2802 of recursive calls to match(). This is different to MATCH_LIMIT because
2803 that limits the total number of calls to match(), not all of which increase
2804 the depth of recursion. Limiting the recursion depth limits the amount of
2805 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2806 when PCRE is compiled, and changed at run time. A patch from Google adds
2807 this functionality to the C++ interface.
2808
2809 18. Changes to the handling of Unicode character properties:
2810
2811 (a) Updated the table to Unicode 4.1.0.
2812
2813 (b) Recognize characters that are not in the table as "Cn" (undefined).
2814
2815 (c) I revised the way the table is implemented to a much improved format
2816 which includes recognition of ranges. It now supports the ranges that
2817 are defined in UnicodeData.txt, and it also amalgamates other
2818 characters into ranges. This has reduced the number of entries in the
2819 table from around 16,000 to around 3,000, thus reducing its size
2820 considerably. I realized I did not need to use a tree structure after
2821 all - a binary chop search is just as efficient. Having reduced the
2822 number of entries, I extended their size from 6 bytes to 8 bytes to
2823 allow for more data.
2824
2825 (d) Added support for Unicode script names via properties such as \p{Han}.
2826
2827 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2828 matching that character.
2829
2830 20. When matching a repeated Unicode property with a minimum greater than zero,
2831 (for example \pL{2,}), PCRE could look past the end of the subject if it
2832 reached it while seeking the minimum number of characters. This could
2833 happen only if some of the characters were more than one byte long, because
2834 there is a check for at least the minimum number of bytes.
2835
2836 21. Refactored the implementation of \p and \P so as to be more general, to
2837 allow for more different types of property in future. This has changed the
2838 compiled form incompatibly. Anybody with saved compiled patterns that use
2839 \p or \P will have to recompile them.
2840
2841 22. Added "Any" and "L&" to the supported property types.
2842
2843 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2844 but give a compile time error if the value is greater than 0xff.
2845
2846 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2847 accidentally not being installed or uninstalled.
2848
2849 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2850 made were to insert the current release number. This seemed silly, because
2851 it made things harder for people building PCRE on systems that don't run
2852 "configure". I have turned pcre.h into a distributed file, no longer built
2853 by "configure", with the version identification directly included. There is
2854 no longer a pcre.h.in file.
2855
2856 However, this change necessitated a change to the pcre-config script as
2857 well. It is built from pcre-config.in, and one of the substitutions was the
2858 release number. I have updated configure.ac so that ./configure now finds
2859 the release number by grepping pcre.h.
2860
2861 26. Added the ability to run the tests under valgrind.
2862
2863
2864 Version 6.4 05-Sep-05
2865 ---------------------
2866
2867 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2868 "--" to be printed when multiple files were scanned, even when none of the
2869 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2870 consider it to be a bug, and have restored the previous behaviour.
2871
2872 2. A couple of code tidies to get rid of compiler warnings.
2873
2874 3. The pcretest program used to cheat by referring to symbols in the library
2875 whose names begin with _pcre_. These are internal symbols that are not
2876 really supposed to be visible externally, and in some environments it is
2877 possible to suppress them. The cheating is now confined to including
2878 certain files from the library's source, which is a bit cleaner.
2879
2880 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2881 file's purpose clearer.
2882
2883 5. Reorganized pcre_ucp_findchar().
2884
2885
2886 Version 6.3 15-Aug-05
2887 ---------------------
2888
2889 1. The file libpcre.pc.in did not have general read permission in the tarball.
2890
2891 2. There were some problems when building without C++ support:
2892
2893 (a) If C++ support was not built, "make install" and "make test" still
2894 tried to test it.
2895
2896 (b) There were problems when the value of CXX was explicitly set. Some
2897 changes have been made to try to fix these, and ...
2898
2899 (c) --disable-cpp can now be used to explicitly disable C++ support.
2900
2901 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2902 backslash in a target when C++ was disabled. This confuses some
2903 versions of "make", apparently. Using an intermediate variable solves
2904 this. (Same for CPP_LOBJ.)
2905
2906 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2907 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2908 necessary on certain architectures.
2909
2910 4. Added a setting of -export-symbols-regex to the link command to remove
2911 those symbols that are exported in the C sense, but actually are local
2912 within the library, and not documented. Their names all begin with
2913 "_pcre_". This is not a perfect job, because (a) we have to except some
2914 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2915 available (and never for static libraries). I have made a note to try to
2916 find a way round (a) in the future.
2917
2918
2919 Version 6.2 01-Aug-05
2920 ---------------------
2921
2922 1. There was no test for integer overflow of quantifier values. A construction
2923 such as {1111111111111111} would give undefined results. What is worse, if
2924 a minimum quantifier for a parenthesized subpattern overflowed and became
2925 negative, the calculation of the memory size went wrong. This could have
2926 led to memory overwriting.
2927
2928 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2929
2930 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2931 operating environments where this matters.
2932
2933 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2934 PCRE options from within the C++ wrapper.
2935
2936 5. Named capturing subpatterns were not being correctly counted when a pattern
2937 was compiled. This caused two problems: (a) If there were more than 100
2938 such subpatterns, the calculation of the memory needed for the whole
2939 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2940 back references of the form \12, where the number was greater than 9, were
2941 not recognized as back references, even though there were sufficient
2942 previous subpatterns.
2943
2944 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2945 versions of gcc, e.g. 2.95.4.
2946
2947
2948 Version 6.1 21-Jun-05
2949 ---------------------
2950
2951 1. There was one reference to the variable "posix" in pcretest.c that was not
2952 surrounded by "#if !defined NOPOSIX".
2953
2954 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2955 the cross-check on the old pcre_info() function, for the benefit of the
2956 cut-down version of PCRE that is currently imported into Exim.
2957
2958 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2959 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2960 patterns that start (?i) but I don't think that matters. The use of (?i) is
2961 just an example; this all applies to the other options as well.
2962
2963 4. Since libtool seems to echo the compile commands it is issuing, the output
2964 from "make" can be reduced a bit by putting "@" in front of each libtool
2965 compile command.
2966
2967 5. Patch from the folks at Google for configure.in to be a bit more thorough
2968 in checking for a suitable C++ installation before trying to compile the
2969 C++ stuff. This should fix a reported problem when a compiler was present,
2970 but no suitable headers.
2971
2972 6. The man pages all had just "PCRE" as their title. I have changed them to
2973 be the relevant file name. I have also arranged that these names are
2974 retained in the file doc/pcre.txt, which is a concatenation in text format
2975 of all the man pages except the little individual ones for each function.
2976
2977 7. The NON-UNIX-USE file had not been updated for the different set of source
2978 files that come with release 6. I also added a few comments about the C++
2979 wrapper.
2980
2981
2982 Version 6.0 07-Jun-05
2983 ---------------------
2984
2985 1. Some minor internal re-organization to help with my DFA experiments.
2986
2987 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2988 didn't matter for the library itself when fully configured, but did matter
2989 when compiling without UCP support, or within Exim, where the ucp files are
2990 not imported.
2991
2992 3. Refactoring of the library code to split up the various functions into
2993 different source modules. The addition of the new DFA matching code (see
2994 below) to a single monolithic source would have made it really too
2995 unwieldy, quite apart from causing all the code to be include in a
2996 statically linked application, when only some functions are used. This is
2997 relevant even without the DFA addition now that patterns can be compiled in
2998 one application and matched in another.
2999
3000 The downside of splitting up is that there have to be some external
3001 functions and data tables that are used internally in different modules of
3002 the library but which are not part of the API. These have all had their
3003 names changed to start with "_pcre_" so that they are unlikely to clash
3004 with other external names.
3005
3006 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3007 a different (DFA) algorithm. Although it is slower than the original
3008 function, it does have some advantages for certain types of matching
3009 problem.
3010
3011 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3012 including restarting after a partial match.
3013
3014 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3015 defined when compiling for Windows was sent to me. I have put it into the
3016 code, though I have no means of testing or verifying it.
3017
3018 7. Added the pcre_refcount() auxiliary function.
3019
3020 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3021 match before or at the first newline in the subject string. In pcretest,
3022 the /f option on a pattern can be used to set this.
3023
3024 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3025 would behave wrongly. This has been present in PCRE since release 4.0.
3026
3027 10. A number of changes to the pcregrep command:
3028
3029 (a) Refactored how -x works; insert ^(...)$ instead of setting
3030 PCRE_ANCHORED and checking the length, in preparation for adding
3031 something similar for -w.
3032
3033 (b) Added the -w (match as a word) option.
3034
3035 (c) Refactored the way lines are read and buffered so as to have more
3036 than one at a time available.
3037
3038 (d) Implemented a pcregrep test script.
3039
3040 (e) Added the -M (multiline match) option. This allows patterns to match
3041 over several lines of the subject. The buffering ensures that at least
3042 8K, or the rest of the document (whichever is the shorter) is available
3043 for matching (and similarly the previous 8K for lookbehind assertions).
3044
3045 (f) Changed the --help output so that it now says
3046
3047 -w, --word-regex(p)
3048
3049 instead of two lines, one with "regex" and the other with "regexp"
3050 because that confused at least one person since the short forms are the
3051 same. (This required a bit of code, as the output is generated
3052 automatically from a table. It wasn't just a text change.)
3053
3054 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3055 option but starts with a hyphen. Could be a pattern or a path name
3056 starting with a hyphen, for instance.
3057
3058 (h) "-" can be given as a file name to represent stdin.
3059
3060 (i) When file names are being printed, "(standard input)" is used for
3061 the standard input, for compatibility with GNU grep. Previously
3062 "<stdin>" was used.
3063
3064 (j) The option --label=xxx can be used to supply a name to be used for
3065 stdin when file names are being printed. There is no short form.
3066
3067 (k) Re-factored the options decoding logic because we are going to add
3068 two more options that take data. Such options can now be given in four
3069 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3070
3071 (l) Added the -A, -B, and -C options for requesting that lines of context
3072 around matches be printed.
3073
3074 (m) Added the -L option to print the names of files that do not contain
3075 any matching lines, that is, the complement of -l.
3076
3077 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3078 continue to scan other files.
3079
3080 (o) The -s option was incorrectly implemented. For compatibility with other
3081 greps, it now suppresses the error message for a non-existent or non-
3082 accessible file (but not the return code). There is a new option called
3083 -q that suppresses the output of matching lines, which was what -s was
3084 previously doing.
3085
3086 (p) Added --include and --exclude options to specify files for inclusion
3087 and exclusion when recursing.
3088
3089 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3090 Hopefully, it now does.
3091
3092 12. Missing cast in pcre_study().
3093
3094 13. Added an "uninstall" target to the makefile.
3095
3096 14. Replaced "extern" in the function prototypes in Makefile.in with
3097 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3098 world, but is set differently for Windows.
3099
3100 15. Added a second compiling function called pcre_compile2(). The only
3101 difference is that it has an extra argument, which is a pointer to an
3102 integer error code. When there is a compile-time failure, this is set
3103 non-zero, in addition to the error test pointer being set to point to an
3104 error message. The new argument may be NULL if no error number is required
3105 (but then you may as well call pcre_compile(), which is now just a
3106 wrapper). This facility is provided because some applications need a
3107 numeric error indication, but it has also enabled me to tidy up the way
3108 compile-time errors are handled in the POSIX wrapper.
3109
3110 16. Added VPATH=.libs to the makefile; this should help when building with one
3111 prefix path and installing with another. (Or so I'm told by someone who
3112 knows more about this stuff than I do.)
3113
3114 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3115 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3116 match everything, including newlines. This is not POSIX-compatible, but
3117 somebody wanted the feature. From pcretest it can be activated by using
3118 both the P and the s flags.
3119
3120 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3121
3122 19. libpcre.pc was being incorrectly installed as executable.
3123
3124 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3125 it now also looks for '\r' so that it will work unmodified on Windows.
3126
3127 21. Added Google's contributed C++ wrapper to the distribution.
3128
3129 22. Added some untidy missing memory free() calls in pcretest, to keep
3130 Electric Fence happy when testing.
3131
3132
3133
3134 Version 5.0 13-Sep-04
3135 ---------------------
3136
3137 1. Internal change: literal characters are no longer packed up into items
3138 containing multiple characters in a single byte-string. Each character
3139 is now matched using a separate opcode. However, there may be more than one
3140 byte in the character in UTF-8 mode.
3141
3142 2. The pcre_callout_block structure has two new fields: pattern_position and
3143 next_item_length. These contain the offset in the pattern to the next match
3144 item, and its length, respectively.
3145
3146 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3147 insertion of callouts before each pattern item. Added the /C option to
3148 pcretest to make use of this.
3149
3150 4. On the advice of a Windows user, the lines
3151
3152 #if defined(_WIN32) || defined(WIN32)
3153 _setmode( _fileno( stdout ), 0x8000 );
3154 #endif /* defined(_WIN32) || defined(WIN32) */
3155
3156 have been added to the source of pcretest. This apparently does useful
3157 magic in relation to line terminators.
3158
3159 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3160 for the benefit of those environments where the "b" makes a difference.
3161
3162 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3163 to know about it. I have put a hack into configure.in that adds in code
3164 to set GCC=yes if CC=icc. This seems to end up at a point in the
3165 generated configure script that is early enough to affect the setting of
3166 compiler options, which is what is needed, but I have no means of testing
3167 whether it really works. (The user who reported this had patched the
3168 generated configure script, which of course I cannot do.)
3169
3170 LATER: After change 22 below (new libtool files), the configure script
3171 seems to know about icc (and also ecc). Therefore, I have commented out
3172 this hack in configure.in.
3173
3174 7. Added support for pkg-config (2 patches were sent in).
3175
3176 8. Negated POSIX character classes that used a combination of internal tables
3177 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3178 [[:^ascii]]. Typically, they would match almost any characters. The other
3179 POSIX classes were not broken in this way.
3180
3181 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3182 to find the match, as PCRE was deluded into thinking that the match had to
3183 start at the start point or following a newline. The same bug applied to
3184 patterns with negative forward assertions or any backward assertions
3185 preceding ".*" at the start, unless the pattern required a fixed first
3186 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3187
3188 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3189 starting at the last subject character, bytes beyond the end of the subject
3190 string were read.
3191
3192 11. Renamed the variable "class" as "classbits" to make life easier for C++
3193 users. (Previously there was a macro definition, but it apparently wasn't
3194 enough.)
3195
3196 12. Added the new field "tables" to the extra data so that tables can be passed
3197 in at exec time, or the internal tables can be re-selected. This allows
3198 a compiled regex to be saved and re-used at a later time by a different
3199 program that might have everything at different addresses.
3200
3201 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3202 -R library as well as a -L library.
3203
3204 14. The debugging options of pcretest (-d on the command line or D on a
3205 pattern) showed incorrect output for anything following an extended class
3206 that contained multibyte characters and which was followed by a quantifier.
3207
3208 15. Added optional support for general category Unicode character properties
3209 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3210 support. It adds about 90K to the size of the library. The meanings of the
3211 inbuilt class escapes such as \d and \s have NOT been changed.
3212
3213 16. Updated pcredemo.c to include calls to free() to release the memory for the
3214 compiled pattern.
3215
3216 17. The generated file chartables.c was being created in the source directory
3217 instead of in the building directory. This caused the build to fail if the
3218 source directory was different from the building directory, and was
3219 read-only.
3220
3221 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3222 file. No doubt somebody will tell me if they don't make sense... Also added
3223 Dan Mooney's comments about building on OpenVMS.
3224
3225 19. Added support for partial matching via the PCRE_PARTIAL option for
3226 pcre_exec() and the \P data escape in pcretest.
3227
3228 20. Extended pcretest with 3 new pattern features:
3229
3230 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3231 write the compiled pattern to the file whose name is "rest-of-line".
3232 This is a straight binary dump of the data, with the saved pointer to
3233 the character tables forced to be NULL. The study data, if any, is
3234 written too. After writing, pcretest reads a new pattern.
3235
3236 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3237 compiled pattern from the given file. There must not be any
3238 occurrences of "<" in the file name (pretty unlikely); if there are,
3239 pcretest will instead treat the initial "<" as a pattern delimiter.
3240 After reading in the pattern, pcretest goes on to read data lines as
3241 usual.
3242
3243 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3244 and 16-bit fields in a compiled pattern, to simulate a pattern that
3245 was compiled on a host of opposite endianness.
3246
3247 21. The pcre-exec() function can now cope with patterns that were compiled on
3248 hosts of opposite endianness, with this restriction:
3249
3250 As for any compiled expression that is saved and used later, the tables
3251 pointer field cannot be preserved; the extra_data field in the arguments
3252 to pcre_exec() should be used to pass in a tables address if a value
3253 other than the default internal tables were used at compile time.
3254
3255 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3256 now diagnosed as an error. Previously, most of the time, a negative number
3257 would have been treated as zero, but if in addition "ovector" was passed as
3258 NULL, a crash could occur.
3259
3260 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3261 new versions from the libtool 1.5 distribution (the last one is a copy of
3262 a file called libtool.m4). This seems to have fixed the need to patch
3263 "configure" to support Darwin 1.3 (which I used to do). However, I still
3264 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3265 workstation).
3266
3267 24. Changed the PCRE licence to be the more standard "BSD" licence.
3268
3269
3270 Version 4.5 01-Dec-03
3271 ---------------------
3272
3273 1. There has been some re-arrangement of the code for the match() function so
3274 that it can be compiled in a version that does not call itself recursively.
3275 Instead, it keeps those local variables that need separate instances for
3276 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3277 needs to "recurse". Keeping track of where control must go is done by means
3278 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3279 hide most of the details from the main code, and operates only if
3280 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3281 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3282 operating.
3283
3284 To make it easier for callers to provide specially tailored get/free
3285 functions for this usage, two new functions, pcre_stack_malloc, and
3286 pcre_stack_free, are used. They are always called in strict stacking order,
3287 and the size of block requested is always the same.
3288
3289 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3290 PCRE has been compiled to use the stack or the heap for recursion. The
3291 -C option of pcretest uses this to show which version is compiled.
3292
3293 A new data escape \S, is added to pcretest; it causes the amounts of store
3294 obtained and freed by both kinds of malloc/free at match time to be added
3295 to the output.
3296
3297 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3298 what's available on my current Linux desktop machine.
3299
3300 3. When matching a UTF-8 string, the test for a valid string at the start has
3301 been extended. If start_offset is not zero, PCRE now checks that it points
3302 to a byte that is the start of a UTF-8 character. If not, it returns
3303 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3304 this is necessary because there may be backward assertions in the pattern.
3305 When matching the same subject several times, it may save resources to use
3306 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3307
3308 4. The code for checking the validity of UTF-8 strings has been tightened so
3309 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3310 containing "overlong sequences".
3311
3312 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3313 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3314 should have been "&", but it just so happened that all the cases this let
3315 through by mistake were picked up later in the function.
3316
3317 6. I had used a variable called "isblank" - this is a C99 function, causing
3318 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3319
3320 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3321 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3322 so the version is shown; (c) stop "make test" echoing "./RunTest".
3323
3324 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3325
3326 9. The prototype for memmove() for systems that don't have it was using
3327 size_t, but the inclusion of the header that defines size_t was later. I've
3328 moved the #includes for the C headers earlier to avoid this.
3329
3330 10. Added some adjustments to the code to make it easier to compiler on certain
3331 special systems:
3332
3333 (a) Some "const" qualifiers were missing.
3334 (b) Added the macro EXPORT before all exported functions; by default this
3335 is defined to be empty.
3336 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3337 that it reads its output file name as an argument instead of writing
3338 to the standard output and assuming this can be redirected.
3339
3340 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3341 class containing characters with values greater than 255, PCRE compilation
3342 went into a loop.
3343
3344 12. A recursive reference to a subpattern that was within another subpattern
3345 that had a minimum quantifier of zero caused PCRE to crash. For example,
3346 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3347 recursion. If the recursively-called subpattern itself had a zero repeat,
3348 that was OK.
3349
3350 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3351 buffer into which it was copied (for escape processing) was still set at
3352 1024, so long lines caused crashes.
3353
3354 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3355 "internal error: code overflow...". This applied to any character class
3356 that was followed by a possessive quantifier.
3357
3358 15. Modified the Makefile to add libpcre.la as a prerequisite for
3359 libpcreposix.la because I was told this is needed for a parallel build to
3360 work.
3361
3362 16. If a pattern that contained .* following optional items at the start was
3363 studied, the wrong optimizing data was generated, leading to matching
3364 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3365 matching string must start with a or b or c. The correct conclusion for
3366 this pattern is that a match can start with any character.
3367
3368
3369 Version 4.4 13-Aug-03
3370 ---------------------
3371
3372 1. In UTF-8 mode, a character class containing characters with values between
3373 127 and 255 was not handled correctly if the compiled pattern was studied.
3374 In fixing this, I have also improved the studying algorithm for such
3375 classes (slightly).
3376
3377 2. Three internal functions had redundant arguments passed to them. Removal
3378 might give a very teeny performance improvement.
3379
3380 3. Documentation bug: the value of the capture_top field in a callout is *one
3381 more than* the number of the hightest numbered captured substring.
3382
3383 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3384 in incorrectly linking with a previously installed version. They now link
3385 explicitly with libpcre.la.
3386
3387 5. configure.in no longer needs to recognize Cygwin specially.
3388
3389 6. A problem in pcre.in for Windows platforms is fixed.
3390
3391 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3392 pcretest, it used to include the size of the study block as part of its
3393 output. Unfortunately, the structure contains a field that has a different
3394 size on different hardware architectures. This meant that the tests that
3395 showed this size failed. As the block is currently always of a fixed size,
3396 this information isn't actually particularly useful in pcretest output, so
3397 I have just removed it.
3398
3399 8. Three pre-processor statements accidentally did not start in column 1.
3400 Sadly, there are *still* compilers around that complain, even though
3401 standard C has not required this for well over a decade. Sigh.
3402
3403 9. In pcretest, the code for checking callouts passed small integers in the
3404 callout_data field, which is a void * field. However, some picky compilers
3405 complained about the casts involved for this on 64-bit systems. Now
3406 pcretest passes the address of the small integer instead, which should get
3407 rid of the warnings.
3408
3409 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3410 both compile and run time, and gives an error if an invalid UTF-8 sequence
3411 is found. There is a option for disabling this check in cases where the
3412 string is known to be correct and/or the maximum performance is wanted.
3413
3414 11. In response to a bug report, I changed one line in Makefile.in from
3415
3416 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3417 to
3418 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3419
3420 to look similar to other lines, but I have no way of telling whether this
3421 is the right thing to do, as I do not use Windows. No doubt I'll get told
3422 if it's wrong...
3423
3424
3425 Version 4.3 21-May-03
3426 ---------------------
3427
3428 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3429 Makefile.
3430
3431 2. Some refactoring to improve the quality of the code:
3432
3433 (i) The utf8_table... variables are now declared "const".
3434
3435 (ii) The code for \cx, which used the "case flipping" table to upper case
3436 lower case letters, now just substracts 32. This is ASCII-specific,
3437 but the whole concept of \cx is ASCII-specific, so it seems
3438 reasonable.
3439
3440 (iii) PCRE was using its character types table to recognize decimal and
3441 hexadecimal digits in the pattern. This is silly, because it handles
3442 only 0-9, a-f, and A-F, but the character types table is locale-
3443 specific, which means strange things might happen. A private
3444 table is now used for this - though it costs 256 bytes, a table is
3445 much faster than multiple explicit tests. Of course, the standard
3446 character types table is still used for matching digits in subject
3447 strings against \d.
3448
3449 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3450 ending in _t are). So I've renamed it as ESC_tee.
3451
3452 3. The first argument for regexec() in the POSIX wrapper should have been
3453 defined as "const".
3454
3455 4. Changed pcretest to use malloc() for its buffers so that they can be
3456 Electric Fenced for debugging.
3457
3458 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3459 to read one or more bytes before the start of the subject string. Often this
3460 had no effect on PCRE's behaviour, but in some circumstances it could
3461 provoke a segmentation fault.
3462
3463 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3464 to try to read one or more bytes before the start of the subject string.
3465
3466 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3467 UTF-8 support could misbehave in various ways if the subject string
3468 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3469 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3470 back over UTF-8 characters.)
3471
3472
3473 Version 4.2 14-Apr-03
3474 ---------------------
3475
3476 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3477
3478 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3479 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3480 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3481 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3482 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3483 and BUILD_EXEEXT
3484 Note: automatic setting of the BUILD variables is not yet working
3485 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3486 compile-time but not at link-time
3487 [LINK]: use for linking executables only
3488 make different versions for Windows and non-Windows
3489 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3490 libraries
3491 [LINK_FOR_BUILD]: new variable
3492 [OBJEXT]: use throughout
3493 [EXEEXT]: use throughout
3494 <winshared>: new target
3495 <wininstall>: new target
3496 <dftables.o>: use native compiler
3497 <dftables>: use native linker
3498 <install>: handle Windows platform correctly
3499 <clean>: ditto
3500 <check>: ditto
3501 copy DLL to top builddir before testing
3502
3503 As part of these changes, -no-undefined was removed again. This was reported
3504 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3505 in any case.
3506
3507 3. Some tidies to get rid of compiler warnings:
3508
3509 . In the match_data structure, match_limit was an unsigned long int, whereas
3510 match_call_count was an int. I've made them both unsigned long ints.
3511
3512 . In pcretest the fact that a const uschar * doesn't automatically cast to
3513 a void * provoked a warning.
3514
3515 . Turning on some more compiler warnings threw up some "shadow" variables
3516 and a few more missing casts.
3517
3518 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3519 option, a class that contained a single character with a value between 128
3520 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3521
3522 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3523 option, a class that contained several characters, but with at least one
3524 whose value was between 128 and 255 caused PCRE to crash.
3525
3526
3527 Version 4.1 12-Mar-03
3528 ---------------------
3529
3530 1. Compiling with gcc -pedantic found a couple of places where casts were
3531 needed, and a string in dftables.c that was longer than standard compilers are
3532 required to support.
3533
3534 2. Compiling with Sun's compiler found a few more places where the code could
3535 be tidied up in order to avoid warnings.
3536
3537 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3538 first of these names is deprecated in the latest Autoconf in favour of the name
3539 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3540 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3541 analogy I have changed it to CFLAGS_FOR_BUILD.
3542
3543 4. Added -no-undefined to the linking command in the Makefile, because this is
3544 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3545 linking step for the pcreposix library.
3546
3547 5. PCRE was failing to diagnose the case of two named groups with the same
3548 name.
3549
3550 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3551 literal character that is needed in the subject for a match, and scans along to
3552 ensure that it is present before embarking on the full matching process. This
3553 saves time in cases of nested unlimited repeats that are never going to match.
3554 Problem: the scan can take a lot of time if the subject is very long (e.g.
3555 megabytes), thus penalizing straightforward matches. It is now done only if the
3556 amount of subject to be scanned is less than 1000 bytes.
3557
3558 7. A lesser problem with the same optimization is that it was recording the
3559 first character of an anchored pattern as "needed", thus provoking a search
3560 right along the subject, even when the first match of the pattern was going to
3561 fail. The "needed" character is now not set for anchored patterns, unless it
3562 follows something in the pattern that is of non-fixed length. Thus, it still
3563 fulfils its original purpose of finding quick non-matches in cases of nested
3564 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3565
3566
3567 Version 4.0 17-Feb-03
3568 ---------------------
3569
3570 1. If a comment in an extended regex that started immediately after a meta-item
3571 extended to the end of string, PCRE compiled incorrect data. This could lead to
3572 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3573
3574 2. Moved to autoconf 2.53 and libtool 1.4.2.
3575
3576 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3577 the special perltest8 script is no longer needed - all the tests can be run
3578 from a single perltest script.
3579
3580 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3581 by \s. It has now been removed in PCRE. This means it isn't recognized as
3582 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3583 class [:space:] *does* include VT, thereby creating a mess.
3584
3585 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3586 space and tab.
3587
3588 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3589 its new features into the main test script, reducing the number of scripts.
3590
3591 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3592 were backward compatible, and made the (?i) apply to the whole pattern, as if
3593 /i were given. Now it behaves more logically, and applies the option setting
3594 only to what follows. PCRE has been changed to follow suit. However, if it
3595 finds options settings right at the start of the pattern, it extracts them into
3596 the global options, as before. Thus, they show up in the info data.
3597
3598 8. Added support for the \Q...\E escape sequence. Characters in between are
3599 treated as literals. This is slightly different from Perl in that $ and @ are
3600 also handled as literals inside the quotes. In Perl, they will cause variable
3601 interpolation. Note the following examples:
3602
3603 Pattern PCRE matches Perl matches
3604
3605 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3606 \Qabc\$xyz\E abc\$xyz abc\$xyz
3607 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3608
3609 For compatibility with Perl, \Q...\E sequences are recognized inside character
3610 classes as well as outside them.
3611
3612 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3613 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3614 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3615 signed/unsigned warnings.
3616
3617 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3618 option for pcretest, so I've replaced it by a simple function that does just
3619 that job.
3620
3621 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3622 "pcregrep -".
3623
3624 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3625 Java package. This provides some syntactic sugar for simple cases of what my
3626 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3627 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3628 item, you can use this simplified notation. Note that only makes sense with
3629 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3630 greediness, whatever the setting of the PCRE_UNGREEDY option.
3631
3632 13. A change of greediness default within a pattern was not taking effect at
3633 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3634 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3635 was abstracted outside.
3636
3637 14. PCRE now supports the \G assertion. It is true when the current matching
3638 position is at the start point of the match. This differs from \A when the
3639 starting offset is non-zero. Used with the /g option of pcretest (or similar
3640 code), it works in the same way as it does for Perl's /g option. If all
3641 alternatives of a regex begin with \G, the expression is anchored to the start
3642 match position, and the "anchored" flag is set in the compiled expression.
3643
3644 15. Some bugs concerning the handling of certain option changes within patterns
3645 have been fixed. These applied to options other than (?ims). For example,
3646 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3647 been the other way round. Some of this was related to change 7 above.
3648
3649 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3650 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3651 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3652 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3653
3654 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3655 mode. Unlike ".", it always matches newline, whatever the setting of
3656 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3657 assertions. Perl allows it, but it doesn't (in general) work because it can't
3658 calculate the length of the lookbehind. At least, that's the case for Perl
3659 5.8.0 - I've been told they are going to document that it doesn't work in
3660 future.
3661
3662 18. Added an error diagnosis for escapes that PCRE does not support: these are
3663 \L, \l, \N, \P, \p, \U, \u, and \X.
3664
3665 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3666 reading past the end of the pattern in cases such as /[abcd/.
3667
3668 20. PCRE was getting more memory than necessary for patterns with classes that
3669 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3670
3671 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3672 compiling PCRE for use with Virtual Pascal.
3673
3674 22. Small fix to the Makefile to make it work properly if the build is done
3675 outside the source tree.
3676
3677 23. Added a new extension: a condition to go with recursion. If a conditional
3678 subpattern starts with (?(R) the "true" branch is used if recursion has
3679 happened, whereas the "false" branch is used only at the top level.
3680
3681 24. When there was a very long string of literal characters (over 255 bytes
3682 without UTF support, over 250 bytes with UTF support), the computation of how
3683 much memory was required could be incorrect, leading to segfaults or other
3684 strange effects.
3685
3686 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3687 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3688 there was a subsequent back reference to those brackets. This meant that, for
3689 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3690 possible to check for precisely this case. All we can do is abandon the
3691 optimization if .* occurs inside capturing brackets when there are any back
3692 references whatsoever. (See below for a better fix that came later.)
3693
3694 26. The handling of the optimization for finding the first character of a
3695 non-anchored pattern, and for finding a character that is required later in the
3696 match were failing in some cases. This didn't break the matching; it just
3697 failed to optimize when it could. The way this is done has been re-implemented.
3698
3699 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3700
3701 28. Added a new feature that provides some of the functionality that Perl
3702 provides with (?{...}). The facility is termed a "callout". The way it is done
3703 in PCRE is for the caller to provide an optional function, by setting
3704 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3705 global variable. By default it is unset, which disables all calling out. To get
3706 the function called, the regex must include (?C) at appropriate points. This
3707 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3708 This provides a means of identifying different callout points. When PCRE
3709 reaches such a point in the regex, if pcre_callout has been set, the external
3710 function is called. It is provided with data in a structure called
3711 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3712 matching continues; if it returns a non-zero value, the match at the current
3713 point fails. However, backtracking will occur if possible. [This was changed
3714 later and other features added - see item 49 below.]
3715
3716 29. pcretest is upgraded to test the callout functionality. It provides a
3717 callout function that displays information. By default, it shows the start of
3718 the match and the current position in the text. There are some new data escapes
3719 to vary what happens:
3720
3721 \C+ in addition, show current contents of captured substrings
3722 \C- do not supply a callout function
3723 \C!n return 1 when callout number n is reached
3724 \C!n!m return 1 when callout number n is reached for the mth time
3725
3726 30. If pcregrep was called with the -l option and just a single file name, it
3727 output "<stdin>" if a match was found, instead of the file name.
3728
3729 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3730 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3731 pcre_exec(). This saves a malloc/free per call. The default value of
3732 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3733 when configuring.
3734
3735 32. The default maximum size of a compiled pattern is 64K. There have been a
3736 few cases of people hitting this limit. The code now uses macros to handle the
3737 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3738 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3739 configuring. Tests 2 and 5 work only with 2-byte links because they output
3740 debugging information about compiled patterns.
3741
3742 33. Internal code re-arrangements:
3743
3744 (a) Moved the debugging function for printing out a compiled regex into
3745 its own source file (printint.c) and used #include to pull it into
3746 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3747 separate copies.
3748
3749 (b) Defined the list of op-code names for debugging as a macro in
3750 internal.h so that it is next to the definition of the opcodes.
3751
3752 (c) Defined a table of op-code lengths for simpler skipping along compiled
3753 code. This is again a macro in internal.h so that it is next to the
3754 definition of the opcodes.
3755
3756 34. Added support for recursive calls to individual subpatterns, along the
3757 lines of Robin Houston's patch (but implemented somewhat differently).
3758
3759 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3760 allow it to read and process whole directories in Win32. This code was
3761 contributed by Lionel Fourquaux; it has not been tested by me.
3762
3763 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3764 used to name a group. Names consist of alphanumerics and underscores, and must
3765 be unique. Back references use the syntax (?P=name) and recursive calls use
3766 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3767 numbers. The function pcre_fullinfo() can be used after compilation to extract
3768 a name/number map. There are three relevant calls:
3769
3770 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3771 PCRE_INFO_NAMECOUNT yields the number of entries
3772 PCRE_INFO_NAMETABLE yields a pointer to the map.
3773
3774 The map is a vector of fixed-size entries. The size of each entry depends on
3775 the length of the longest name used. The first two bytes of each entry are the
3776 group number, most significant byte first. There follows the corresponding
3777 name, zero terminated. The names are in alphabetical order.
3778
3779 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3780 case instead of 255. Making it the same both with and without UTF-8 support
3781 means that the same test output works with both.
3782
3783 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3784 calling malloc() with a zero argument.
3785
3786 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3787 optimization. I've improved things by keeping a bitmap of backreferences with
3788 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3789 fact referenced, the optimization can be applied. It is unlikely that a
3790 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3791 the match to follow \n) will appear inside brackets with a number greater than
3792 31, but if it does, any back reference > 31 suppresses the optimization.
3793
3794 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3795 of disabling numbered capturing parentheses. Any opening parenthesis that is
3796 not followed by ? behaves as if it were followed by ?: but named parentheses
3797 can still be used for capturing (and they will acquire numbers in the usual
3798 way).
3799
3800 41. Redesigned the return codes from the match() function into yes/no/error so
3801 that errors can be passed back from deep inside the nested calls. A malloc
3802 failure while inside a recursive subpattern call now causes the
3803 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3804
3805 42. It is now possible to set a limit on the number of times the match()
3806 function is called in a call to pcre_exec(). This facility makes it possible to
3807 limit the amount of recursion and backtracking, though not in a directly
3808 obvious way, because the match() function is used in a number of different
3809 circumstances. The count starts from zero for each position in the subject
3810 string (for non-anchored patterns). The default limit is, for compatibility, a
3811 large number, namely 10 000 000. You can change this in two ways:
3812
3813 (a) When configuring PCRE before making, you can use --with-match-limit=n
3814 to set a default value for the compiled library.
3815
3816 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3817 a different value is set. See 45 below.
3818
3819 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3820
3821 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3822 of things that can be changed at compile time. The first argument specifies
3823 what is wanted and the second points to where the information is to be placed.
3824 The current list of available information is:
3825
3826 PCRE_CONFIG_UTF8
3827
3828 The output is an integer that is set to one if UTF-8 support is available;
3829 otherwise it is set to zero.
3830
3831 PCRE_CONFIG_NEWLINE
3832
3833 The output is an integer that it set to the value of the code that is used for
3834 newline. It is either LF (10) or CR (13).
3835
3836 PCRE_CONFIG_LINK_SIZE
3837
3838 The output is an integer that contains the number of bytes used for internal
3839 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3840
3841 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3842
3843 The output is an integer that contains the threshold above which the POSIX
3844 interface uses malloc() for output vectors. See item 31 above.
3845
3846 PCRE_CONFIG_MATCH_LIMIT
3847
3848 The output is an unsigned integer that contains the default limit of the number
3849 of match() calls in a pcre_exec() execution. See 42 above.
3850
3851 44. pcretest has been upgraded by the addition of the -C option. This causes it
3852 to extract all the available output from the new pcre_config() function, and to
3853 output it. The program then exits immediately.
3854
3855 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3856 order to support additional features. One way would have been to define
3857 pcre_exec2() (for example) with extra arguments, but this would not have been
3858 extensible, and would also have required all calls to the original function to
3859 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3860 is used for passing in "extra" data from pcre_study().
3861
3862 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3863 contains the following fields:
3864
3865 flags a bitmap indicating which of the following fields are set
3866 study_data opaque data from pcre_study()
3867 match_limit a way of specifying a limit on match() calls for a specific
3868 call to pcre_exec()
3869 callout_data data for callouts (see 49 below)
3870
3871 The flag bits are also defined in pcre.h, and are
3872
3873 PCRE_EXTRA_STUDY_DATA
3874 PCRE_EXTRA_MATCH_LIMIT
3875 PCRE_EXTRA_CALLOUT_DATA
3876
3877 The pcre_study() function now returns one of these new pcre_extra blocks, with
3878 the actual study data pointed to by the study_data field, and the
3879 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3880 before. That is, this change is entirely upwards-compatible and requires no
3881 change to existing code.
3882
3883 If you want to pass in additional data to pcre_exec(), you can either place it
3884 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3885 block.
3886
3887 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3888 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3889 times with different match limits, until it finds the minimum value needed for
3890 pcre_exec() to complete. The value is then output. This can be instructive; for
3891 most simple matches the number is quite small, but for pathological cases it
3892 gets very large very quickly.
3893
3894 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3895 returns the size of the data block pointed to by the study_data field in a
3896 pcre_extra block, that is, the value that was passed as the argument to
3897 pcre_malloc() when PCRE was getting memory in which to place the information
3898 created by pcre_study(). The fourth argument should point to a size_t variable.
3899 pcretest has been extended so that this information is shown after a successful
3900 pcre_study() call when information about the compiled regex is being displayed.
3901
3902 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3903 because what follows is always an absolute path. (Later: it turns out that this
3904 is more than cosmetic for MinGW, because it doesn't like empty path
3905 components.)
3906
3907 49. Some changes have been made to the callout feature (see 28 above):
3908
3909 (i) A callout function now has three choices for what it returns:
3910
3911 0 => success, carry on matching
3912 > 0 => failure at this point, but backtrack if possible
3913 < 0 => serious error, return this value from pcre_exec()
3914
3915 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3916 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3917 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3918 use by callout functions. It will never be used by PCRE itself.
3919
3920 (ii) The pcre_extra structure (see 45 above) has a void * field called
3921 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3922 pcre_callout_block structure has a field of the same name. The contents of
3923 the field passed in the pcre_extra structure are passed to the callout
3924 function in the corresponding field in the callout block. This makes it
3925 easier to use the same callout-containing regex from multiple threads. For
3926 testing, the pcretest program has a new data escape
3927
3928 \C*n pass the number n (may be negative) as callout_data
3929
3930 If the callout function in pcretest receives a non-zero value as
3931 callout_data, it returns that value.
3932
3933 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3934 there were some redundant $(CFLAGS) in commands that are now specified as
3935 $(LINK), which already includes $(CFLAGS).
3936
3937 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3938 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3939 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3940 one-byte characters throughout. Note that case-insensitive matching applies
3941 only to characters whose values are less than 256. PCRE doesn't support the
3942 notion of cases for higher-valued characters.
3943
3944 (i) A character class whose characters are all within 0-255 is handled as
3945 a bit map, and the map is inverted for negative classes. Previously, a
3946 character > 255 always failed to match such a class; however it should
3947 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3948
3949 (ii) A negated character class with a single character < 255 is coded as
3950 "not this character" (OP_NOT). This wasn't working properly when the test
3951 character was multibyte, either singly or repeated.
3952
3953 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3954 mode, for example: \x{100}{2,3}.
3955
3956 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3957 singly or repeated) now correctly test multibyte characters. However,
3958 PCRE doesn't recognize any characters with values greater than 255 as
3959 digits, spaces, or word characters. Such characters always match \D, \S,
3960 and \W, and never match \d, \s, or \w.
3961
3962 (v) Classes may now contain characters and character ranges with values
3963 greater than 255. For example: [ab\x{100}-\x{400}].
3964
3965 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3966 PCRE in UTF-8 mode.
3967
3968 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3969 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3970 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3971 value.)
3972
3973 53. The single man page has become too large. I have therefore split it up into
3974 a number of separate man pages. These also give rise to individual HTML pages;
3975 these are now put in a separate directory, and there is an index.html page that
3976 lists them all. Some hyperlinking between the pages has been installed.
3977
3978 54. Added convenience functions for handling named capturing parentheses.
3979
3980 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3981 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3982 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3983 are faulted.
3984
3985 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3986 calling configure. These values are used when compiling the dftables.c program
3987 which is run to generate the source of the default character tables. They
3988 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3989 you will need to set these values.
3990
3991 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3992
3993
3994 Version 3.9 02-Jan-02
3995 ---------------------
3996
3997 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3998
3999 2. If --disable-static was given, the building process failed when trying to
4000 build pcretest and pcregrep. (For some reason it was using libtool to compile
4001 them, which is not right, as they aren't part of the library.)
4002
4003
4004 Version 3.8 18-Dec-01
4005 ---------------------
4006
4007 1. The experimental UTF-8 code was completely screwed up. It was packing the
4008 bytes in the wrong order. How dumb can you get?
4009
4010
4011 Version 3.7 29-Oct-01
4012 ---------------------
4013
4014 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4015 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4016 this didn't happen under Solaris 8, where I normally test things.
4017
4018 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4019 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4020 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4021 this fix an hour or so after the initial 3.7 release.)
4022
4023
4024 Version 3.6 23-Oct-01
4025 ---------------------
4026
4027 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4028 offsets passed as NULL with zero offset count.
4029
4030 2. The config.guess and config.sub files had not been updated when I moved to
4031 the latest autoconf.
4032
4033
4034 Version 3.5 15-Aug-01
4035 ---------------------
4036
4037 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4038 had been forgotten.
4039
4040 2. By using declared but undefined structures, we can avoid using "void"
4041 definitions in pcre.h while keeping the internal definitions of the structures
4042 private.
4043
4044 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4045 user point of view, this means that both static and shared libraries are built
4046 by default, but this can be individually controlled. More of the work of
4047 handling this static/shared cases is now inside libtool instead of PCRE's make
4048 file.
4049
4050 4. The pcretest utility is now installed along with pcregrep because it is
4051 useful for users (to test regexs) and by doing this, it automatically gets
4052 relinked by libtool. The documentation has been turned into a man page, so
4053 there are now .1, .txt, and .html versions in /doc.
4054
4055 5. Upgrades to pcregrep:
4056 (i) Added long-form option names like gnu grep.
4057 (ii) Added --help to list all options with an explanatory phrase.
4058 (iii) Added -r, --recursive to recurse into sub-directories.
4059 (iv) Added -f, --file to read patterns from a file.
4060
4061 6. pcre_exec() was referring to its "code" argument before testing that
4062 argument for NULL (and giving an error if it was NULL).
4063
4064 7. Upgraded Makefile.in to allow for compiling in a different directory from
4065 the source directory.
4066
4067 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4068 options bits, the pointer it was passed was to an int instead of to an unsigned
4069 long int. This mattered only on 64-bit systems.
4070
4071 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4072 generated) instead of pcre.in, which it its source. Also made the same change
4073 in several of the .c files.
4074
4075 10. A new release of gcc defines printf() as a macro, which broke pcretest
4076 because it had an ifdef in the middle of a string argument for printf(). Fixed
4077 by using separate calls to printf().
4078
4079 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4080 script, to force use of CR or LF instead of \n in the source. On non-Unix
4081 systems, the value can be set in config.h.
4082
4083 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4084 absolute limit. Changed the text of the error message to make this clear, and
4085 likewise updated the man page.
4086
4087 13. The limit of 99 on the number of capturing subpatterns has been removed.
4088 The new limit is 65535, which I hope will not be a "real" limit.
4089
4090
4091 Version 3.4 22-Aug-00
4092 ---------------------
4093
4094 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4095
4096 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4097
4098
4099 Version 3.3 01-Aug-00
4100 ---------------------
4101
4102 1. If an octal character was given, but the value was greater than \377, it
4103 was not getting masked to the least significant bits, as documented. This could
4104 lead to crashes in some systems.
4105
4106 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4107 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4108
4109 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4110 These just pass their arguments on to (pcre_free)(), but they are provided
4111 because some uses of PCRE bind it to non-C systems that can call its functions,
4112 but cannot call free() or pcre_free() directly.
4113
4114 4. Add "make test" as a synonym for "make check". Corrected some comments in
4115 the Makefile.
4116
4117 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4118 Makefile.
4119
4120 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4121 command called pgrep for grepping around the active processes.
4122
4123 7. Added the beginnings of support for UTF-8 character strings.
4124
4125 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4126 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4127 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4128 out for the ar command.)
4129
4130
4131 Version 3.2 12-May-00
4132 ---------------------
4133
4134 This is purely a bug fixing release.
4135
4136 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4137 of ZA. This was just one example of several cases that could provoke this bug,
4138 which was introduced by change 9 of version 2.00. The code for breaking
4139 infinite loops after an iteration that matches an empty string was't working
4140 correctly.
4141
4142 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4143 when matched against abbab (for example). After matching an empty string, it
4144 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4145 caused it to match further down the string than it should.
4146
4147 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4148 was there because it doesn't seem to be needed, and it causes trouble on some
4149 systems, as it is not a Standard C header. It has been removed.
4150
4151 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4152 were reported on the Macintosh. The changes were from
4153
4154 while ((c = *(++ptr)) != 0 && c != '\n');
4155 to
4156 while ((c = *(++ptr)) != 0 && c != '\n') ;
4157
4158 Totally extraordinary, but if that's what it takes...
4159
4160 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4161 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4162 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4163 assumes the way PCRE uses memmove() (always moving upwards).
4164
4165 6. PCRE is being used in one environment where strchr() is not available. There
4166 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4167 faster code anyway.
4168
4169
4170 Version 3.1 09-Feb-00
4171 ---------------------
4172
4173 The only change in this release is the fixing of some bugs in Makefile.in for
4174 the "install" target:
4175
4176 (1) It was failing to install pcreposix.h.
4177
4178 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4179
4180
4181 Version 3.0 01-Feb-00
4182 ---------------------
4183
4184 1. Add support for the /+ modifier to perltest (to output $` like it does in
4185 pcretest).
4186
4187 2. Add support for the /g modifier to perltest.
4188
4189 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4190 matches null strings.
4191
4192 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4193 pattern. Perl treats empty patterns specially - it reuses the most recent
4194 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4195 effect.
4196
4197 5. The POSIX interface was broken in that it was just handing over the POSIX
4198 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4199 required a bigger vector, with some working space on the end. This means that
4200 the POSIX wrapper now has to get and free some memory, and copy the results.
4201
4202 6. Added some simple autoconf support, placing the test data and the
4203 documentation in separate directories, re-organizing some of the
4204 information files, and making it build pcre-config (a GNU standard). Also added
4205 libtool support for building PCRE as a shared library, which is now the
4206 default.
4207
4208 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4209 09 are not valid octal constants. Single digits will be used for minor values
4210 less than 10.
4211
4212 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4213 existing programs that set these in the POSIX interface can use PCRE without
4214 modification.
4215
4216 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4217 return all that pcre_info() returns, plus additional data. The pcre_info()
4218 function is retained for compatibility, but is considered to be obsolete.
4219
4220 10. Added experimental recursion feature (?R) to handle one common case that
4221 Perl 5.6 will be able to do with (?p{...}).
4222
4223 11. Added support for POSIX character classes like [:alpha:], which Perl is
4224 adopting.
4225
4226
4227 Version 2.08 31-Aug-99
4228 ----------------------
4229
4230 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4231 trying to match at the startoffset position, but instead was moving forward to
4232 the next newline as if a previous match had failed.
4233
4234 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4235 and could get into a loop if a null string was matched other than at the start
4236 of the subject.
4237
4238 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4239 be distinguished at compile time, and for completeness also added PCRE_DATE.
4240
4241 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4242 in GnuWin32 environments.
4243
4244
4245 Version 2.07 29-Jul-99
4246 ----------------------
4247
4248 1. The documentation is now supplied in plain text form and HTML as well as in
4249 the form of man page sources.
4250
4251 2. C++ compilers don't like assigning (void *) values to other pointer types.
4252 In particular this affects malloc(). Although there is no problem in Standard
4253 C, I've put in casts to keep C++ compilers happy.
4254
4255 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4256 should be (const char *).
4257
4258 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4259 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4260 However, I haven't made this a standard facility. The documentation doesn't
4261 mention it, and the Makefile doesn't support it.
4262
4263 5. The Makefile now contains an "install" target, with editable destinations at
4264 the top of the file. The pcretest program is not installed.
4265
4266 6. pgrep -V now gives the PCRE version number and date.
4267
4268 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4269 causing the entire string to be ignored, instead of just the last character.
4270
4271 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4272 non-matching string, it can take a very, very long time, even for strings of
4273 quite modest length, because of the nested recursion. PCRE now does better in
4274 some of these cases. It does this by remembering the last required literal
4275 character in the pattern, and pre-searching the subject to ensure it is present
4276 before running the real match. In other words, it applies a heuristic to detect
4277 some types of certain failure quickly, and in the above example, if presented
4278 with a string that has no trailing " it gives "no match" very quickly.
4279
4280 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4281 other alternatives are tried instead.
4282
4283
4284 Version 2.06 09-Jun-99
4285 ----------------------
4286
4287 1. Change pcretest's output for amount of store used to show just the code
4288 space, because the remainder (the data block) varies in size between 32-bit and
4289 64-bit systems.
4290
4291 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4292 start matching at. This allows lookbehinds to work when searching for multiple
4293 occurrences in a string.
4294
4295 3. Added additional options to pcretest for testing multiple occurrences:
4296
4297 /+ outputs the rest of the string that follows a match
4298 /g loops for multiple occurrences, using the new startoffset argument
4299 /G loops for multiple occurrences by passing an incremented pointer
4300
4301 4. PCRE wasn't doing the "first character" optimization for patterns starting
4302 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4303 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4304 the letter 'x'. On long subject strings, this gives a significant speed-up.
4305
4306
4307 Version 2.05 21-Apr-99
4308 ----------------------
4309
4310 1. Changed the type of magic_number from int to long int so that it works
4311 properly on 16-bit systems.
4312
4313 2. Fixed a bug which caused patterns starting with .* not to work correctly
4314 when the subject string contained newline characters. PCRE was assuming
4315 anchoring for such patterns in all cases, which is not correct because .* will
4316 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4317 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4318 must be retried after every newline in the subject.
4319
4320
4321 Version 2.04 18-Feb-99
4322 ----------------------
4323
4324 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4325 computation of the store needed to hold the pattern was incorrect (too large).
4326 If such patterns were nested a few deep, this could multiply and become a real
4327 problem.
4328
4329 2. Added /M option to pcretest to show the memory requirement of a specific
4330 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4331
4332 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4333 compiled in such a way that the backtracking after subsequent failure was
4334 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4335 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4336
4337
4338 Version 2.03 02-Feb-99
4339 ----------------------
4340
4341 1. Fixed typo and small mistake in man page.
4342
4343 2. Added 4th condition (GPL supersedes if conflict) and created separate
4344 LICENCE file containing the conditions.
4345
4346 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4347 Perl, that is the internal \ allows the delimiter to be included in the
4348 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4349 the final delimiter, add \ to the end of the pattern (to test the error).
4350
4351 4. Added the convenience functions for extracting substrings after a successful
4352 match. Updated pcretest to make it able to test these functions.
4353
4354
4355 Version 2.02 14-Jan-99
4356 ----------------------
4357
4358 1. Initialized the working variables associated with each extraction so that
4359 their saving and restoring doesn't refer to uninitialized store.
4360
4361 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4362 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4363 fix the problem.
4364
4365 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4366 calls, and wasn't printing the correct value for compiling calls. Increased the
4367 default value of LOOPREPEAT, and the number of significant figures in the
4368 times.
4369
4370 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4371
4372 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4373 a building problem on Windows NT with a FAT file system.
4374
4375
4376 Version 2.01 21-Oct-98
4377 ----------------------
4378
4379 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4380 to character tables built by pcre_maketables() in the current locale. If NULL
4381 is passed, the default tables are used.
4382
4383
4384 Version 2.00 24-Sep-98
4385 ----------------------
4386
4387 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4388 it any more.
4389
4390 2. Allow quantification of (?>) groups, and make it work correctly.
4391
4392 3. The first character computation wasn't working for (?>) groups.
4393
4394 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4395 end of the subject) and add 5.005's \z, which really does match only at the
4396 very end of the subject.
4397
4398 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4399
4400 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4401 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4402 localized options. All options to pcre_study() were also removed.
4403
4404 7. Add other new features from 5.005:
4405
4406 $(?<= positive lookbehind
4407 $(?<! negative lookbehind
4408 (?imsx-imsx) added the unsetting capability
4409 such a setting is global if at outer level; local otherwise
4410 (?imsx-imsx:) non-capturing groups with option setting
4411 (?(cond)re|re) conditional pattern matching
4412
4413 A backreference to itself in a repeated group matches the previous
4414 captured string.
4415
4416 8. General tidying up of studying (both automatic and via "study")
4417 consequential on the addition of new assertions.
4418
4419 9. As in 5.005, unlimited repeated groups that could match an empty substring
4420 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4421 runtime if any iteration does actually match an empty substring.
4422
4423 10. Include the RunTest script in the distribution.
4424
4425 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4426 discrepancies, some of which were old and were also with respect to 5.004. They
4427 have now been fixed.
4428
4429
4430 Version 1.09 28-Apr-98
4431 ----------------------
4432
4433 1. A negated single character class followed by a quantifier with a minimum
4434 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4435 program crashes, or just wrong answers. This did not apply to negated classes
4436 containing more than one character, or to minima other than one.
4437
4438
4439 Version 1.08 27-Mar-98
4440 ----------------------
4441
4442 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4443
4444 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4445 latter must appear before anything that relies on it in the pattern.
4446
4447
4448 Version 1.07 16-Feb-98
4449 ----------------------
4450
4451 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4452 repeat of a potentially empty string).
4453
4454
4455 Version 1.06 23-Jan-98
4456 ----------------------
4457
4458 1. Added Markus Oberhumer's little patches for C++.
4459
4460 2. Literal strings longer than 255 characters were broken.
4461
4462
4463 Version 1.05 23-Dec-97
4464 ----------------------
4465
4466 1. Negated character classes containing more than one character were failing if
4467 PCRE_CASELESS was set at run time.
4468
4469
4470 Version 1.04 19-Dec-97
4471 ----------------------
4472
4473 1. Corrected the man page, where some "const" qualifiers had been omitted.
4474
4475 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4476 input syntax.
4477
4478 3. Fixed memory leak which occurred when a regex with back references was
4479 matched with an offsets vector that wasn't big enough. The temporary memory
4480 that is used in this case wasn't being freed if the match failed.
4481
4482 4. Tidied pcretest to ensure it frees memory that it gets.
4483
4484 5. Temporary memory was being obtained in the case where the passed offsets
4485 vector was exactly big enough.
4486
4487 6. Corrected definition of offsetof() from change 5 below.
4488
4489 7. I had screwed up change 6 below and broken the rules for the use of
4490 setjmp(). Now fixed.
4491
4492
4493 Version 1.03 18-Dec-97
4494 ----------------------
4495
4496 1. A erroneous regex with a missing opening parenthesis was correctly
4497 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4498 on some systems.
4499
4500 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4501 it was reported that one broken compiler failed on the former because "code" is
4502 also an independent variable.
4503
4504 3. The erroneous regex a[]b caused an array overrun reference.
4505
4506 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4507 fail on data ending with that character. (It was going on too far, and checking
4508 the next character, typically a binary zero.) This was specific to the
4509 optimized code for single-character negative classes.
4510
4511 5. Added a contributed patch from the TIN world which does the following:
4512
4513 + Add an undef for memmove, in case the the system defines a macro for it.
4514
4515 + Add a definition of offsetof(), in case there isn't one. (I don't know
4516 the reason behind this - offsetof() is part of the ANSI standard - but
4517 it does no harm).
4518
4519 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4520 most of the places where whitespace preceded '#'. I have given up and
4521 allowed the remaining 2 cases to be at the margin.
4522
4523 + Rename some variables in pcre to eliminate shadowing. This seems very
4524 pedantic, but does no harm, of course.
4525
4526 6. Moved the call to setjmp() into its own function, to get rid of warnings
4527 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4528
4529 7. Constructs such as \d{8,} were compiling into the equivalent of
4530 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4531 outcome, but in this particular case used more store than had been allocated,
4532 which caused the bug to be discovered because it threw up an internal error.
4533
4534 8. The debugging code in both pcre and pcretest for outputting the compiled
4535 form of a regex was going wrong in the case of back references followed by
4536 curly-bracketed repeats.
4537
4538
4539 Version 1.02 12-Dec-97
4540 ----------------------
4541
4542 1. Typos in pcre.3 and comments in the source fixed.
4543
4544 2. Applied a contributed patch to get rid of places where it used to remove
4545 'const' from variables, and fixed some signed/unsigned and uninitialized
4546 variable warnings.
4547
4548 3. Added the "runtest" target to Makefile.
4549
4550 4. Set default compiler flag to -O2 rather than just -O.
4551
4552
4553 Version 1.01 19-Nov-97
4554 ----------------------
4555
4556 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4557 like /([ab]*)*/, that is, for classes with more than one character in them.
4558
4559 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4560 as /((?>a*))*/ (a PCRE_EXTRA facility).
4561
4562
4563 Version 1.00 18-Nov-97
4564 ----------------------
4565
4566 1. Added compile-time macros to support systems such as SunOS4 which don't have
4567 memmove() or strerror() but have other things that can be used instead.
4568
4569 2. Arranged that "make clean" removes the executables.
4570
4571
4572 Version 0.99 27-Oct-97
4573 ----------------------
4574
4575 1. Fixed bug in code for optimizing classes with only one character. It was
4576 initializing a 32-byte map regardless, which could cause it to run off the end
4577 of the memory it had got.
4578
4579 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4580
4581
4582 Version 0.98 22-Oct-97
4583 ----------------------
4584
4585 1. Fixed bug in code for handling temporary memory usage when there are more
4586 back references than supplied space in the ovector. This could cause segfaults.
4587
4588
4589 Version 0.97 21-Oct-97
4590 ----------------------
4591
4592 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4593
4594 2. Optimized negated single characters not to use a bit map.
4595
4596 3. Brought error texts together as macro definitions; clarified some of them;
4597 fixed one that was wrong - it said "range out of order" when it meant "invalid
4598 escape sequence".
4599
4600 4. Changed some char * arguments to const char *.
4601
4602 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4603
4604 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4605 pcretest.
4606
4607
4608 Version 0.96 16-Oct-97
4609 ----------------------
4610
4611 1. Added a simple "pgrep" utility to the distribution.
4612
4613 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4614 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4615 where "ddd" means "one or more decimal digits".
4616
4617 3. Fixed serious bug. If a pattern had a back reference, but the call to
4618 pcre_exec() didn't supply a large enough ovector to record the related
4619 identifying subpattern, the match always failed. PCRE now remembers the number
4620 of the largest back reference, and gets some temporary memory in which to save
4621 the offsets during matching if necessary, in order to ensure that
4622 backreferences always work.
4623
4624 4. Increased the compatibility with Perl in a number of ways:
4625
4626 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4627 to request this handling. The option can be set at compile or exec time.
4628
4629 (b) $ matches before a terminating newline by default; an option
4630 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4631 mode). The option can be set at compile or exec time.
4632
4633 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4634 the same as Perl's. If the decimal number it represents is less than 10
4635 or there aren't that many previous left capturing parentheses, an octal
4636 escape is read. Inside a character class, it's always an octal escape,
4637 even if it is a single digit.
4638
4639 (d) An escaped but undefined alphabetic character is taken as a literal,
4640 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4641 escapes.
4642
4643 (e) {0} is now permitted. (The previous item is removed from the compiled
4644 pattern).
4645
4646 5. Changed all the names of code files so that the basic parts are no longer
4647 than 10 characters, and abolished the teeny "globals.c" file.
4648
4649 6. Changed the handling of character classes; they are now done with a 32-byte
4650 bit map always.
4651
4652 7. Added the -d and /D options to pcretest to make it possible to look at the
4653 internals of compilation without having to recompile pcre.
4654
4655
4656 Version 0.95 23-Sep-97
4657 ----------------------
4658
4659 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4660 \x20 at the start of a run of normal characters. These were being treated as
4661 real characters, instead of the source characters being re-checked.
4662
4663
4664 Version 0.94 18-Sep-97
4665 ----------------------
4666
4667 1. The functions are now thread-safe, with the caveat that the global variables
4668 containing pointers to malloc() and free() or alternative functions are the
4669 same for all threads.
4670
4671 2. Get pcre_study() to generate a bitmap of initial characters for non-
4672 anchored patterns when this is possible, and use it if passed to pcre_exec().
4673
4674
4675 Version 0.93 15-Sep-97
4676 ----------------------
4677
4678 1. /(b)|(:+)/ was computing an incorrect first character.
4679
4680 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4681 but not actually doing anything yet.
4682
4683 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4684 as Perl does (e.g. [-az] or [az-]).
4685
4686 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4687 all possible positions.
4688
4689 5. Split up into different modules to avoid including unneeded functions in a
4690 compiled binary. However, compile and exec are still in one module. The "study"
4691 function is split off.
4692
4693 6. The character tables are now in a separate module whose source is generated
4694 by an auxiliary program - but can then be edited by hand if required. There are
4695 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4696 toupper() in the code.
4697
4698 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4699 make them global. Abolish the function for setting them, as the caller can now
4700 set them directly.
4701
4702
4703 Version 0.92 11-Sep-97
4704 ----------------------
4705
4706 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4707 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4708
4709 2. Caseless matching was not working in character classes if the characters in
4710 the pattern were in upper case.
4711
4712 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4713
4714 4. Make PCRE_ANCHORED public and accept as a compile option.
4715
4716 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4717 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4718 pass them.
4719
4720 6. Give an error if bad option bits passed at compile or run time.
4721
4722 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4723 pcretest to cause it to pass that flag.
4724
4725 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4726 options, and the first character, if set.
4727
4728 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4729
4730
4731 Version 0.91 10-Sep-97
4732 ----------------------
4733
4734 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4735 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4736
4737 2. PCRE was looping on encountering an indefinitely repeated back reference to
4738 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4739 Perl does - treats the match as successful.
4740
4741 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5