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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 933 - (show annotations)
Sat Feb 25 12:18:23 2012 UTC (7 years, 8 months ago) by ph10
File size: 228141 byte(s)
Error occurred while calculating annotation data.
Applied Graycode's patch to use heap stack frames more efficiently.
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
59 Version 8.30 04-February-2012
60 -----------------------------
61
62 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
63 name is defined in ctype.h.
64
65 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
66 only in quite long subpatterns.
67
68 3. Removed the function pcre_info(), which has been obsolete and deprecated
69 since it was replaced by pcre_fullinfo() in February 2000.
70
71 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
72 match a (*MARK), and the match failed at the start of the subject, a
73 reference to memory before the start of the subject could occur. This bug
74 was introduced by fix 17 of release 8.21.
75
76 5. A reference to an unset group with zero minimum repetition was giving
77 totally wrong answers (in non-JavaScript-compatibility mode). For example,
78 /(another)?(\1?)test/ matched against "hello world test". This bug was
79 introduced in release 8.13.
80
81 6. Add support for 16-bit character strings (a large amount of work involving
82 many changes and refactorings).
83
84 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
85 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
86 from a file.
87
88 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
89 rounding is not applied in this particular case).
90
91 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
92 if they appear, or are escaped, in patterns.
93
94 10. Get rid of a number of -Wunused-but-set-variable warnings.
95
96 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
97 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
98 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
99 also returns the mark "x". This bug applied to capturing parentheses,
100 non-capturing parentheses, and atomic parentheses. It also applied to some
101 assertions.
102
103 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
104 information out of configure.ac instead of relying on pcre.h.generic, which
105 is not stored in the repository.
106
107 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
108 -lreadline.
109
110 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
111
112 15. Applied Graycode's patch to put the top-level frame on the stack rather
113 than the heap when not using the stack for recursion. This gives a
114 performance improvement in many cases when recursion is not deep.
115
116 16. Experimental code added to "pcretest -C" to output the stack frame size.
117
118
119 Version 8.21 12-Dec-2011
120 ------------------------
121
122 1. Updating the JIT compiler.
123
124 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
125 are added as well.
126
127 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
128 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
129 calling _pcre_jit_exec. Some extra comments are added.
130
131 4. (*MARK) settings inside atomic groups that do not contain any capturing
132 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
133 was introduced by change 18 for 8.20.
134
135 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
136 ECMA-262 standard.
137
138 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
139 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
140 This bug was probably introduced by change 9 of 8.13.
141
142 7. While fixing 6 above, I noticed that a number of other items were being
143 incorrectly rejected as "not fixed length". This arose partly because newer
144 opcodes had not been added to the fixed-length checking code. I have (a)
145 corrected the bug and added tests for these items, and (b) arranged for an
146 error to occur if an unknown opcode is encountered while checking for fixed
147 length instead of just assuming "not fixed length". The items that were
148 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
149 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
150 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
151
152 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
153 being incorrectly compiled and would have given unpredicatble results.
154
155 9. A possessively repeated subpattern with minimum repeat count greater than
156 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
157 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
158 the first (A) could occur when it should not.
159
160 10. Add a cast and remove a redundant test from the code.
161
162 11. JIT should use pcre_malloc/pcre_free for allocation.
163
164 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
165 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
166 is anything other than /usr, -L is still shown).
167
168 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
169
170 14. Perl does not support \N without a following name in a [] class; PCRE now
171 also gives an error.
172
173 15. If a forward reference was repeated with an upper limit of around 2000,
174 it caused the error "internal error: overran compiling workspace". The
175 maximum number of forward references (including repeats) was limited by the
176 internal workspace, and dependent on the LINK_SIZE. The code has been
177 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
178 the default depends on LINK_SIZE. There is a new upper limit (for safety)
179 of around 200,000 forward references. While doing this, I also speeded up
180 the filling in of repeated forward references.
181
182 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
183 incorrectly expecting the subject to contain another "a" after the start.
184
185 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
186 in the match, the SKIP should be ignored. This was not happening; instead
187 the SKIP was being treated as NOMATCH. For patterns such as
188 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
189 tested.
190
191 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
192 now much more compatible with Perl, in particular in cases where the result
193 is a non-match for a non-anchored pattern. For example, if
194 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
195 "m", where previously it did not return a name. A side effect of this
196 change is that for partial matches, the last encountered mark name is
197 returned, as for non matches. A number of tests that were previously not
198 Perl-compatible have been moved into the Perl-compatible test files. The
199 refactoring has had the pleasing side effect of removing one argument from
200 the match() function, thus reducing its stack requirements.
201
202 19. If the /S+ option was used in pcretest to study a pattern using JIT,
203 subsequent uses of /S (without +) incorrectly behaved like /S+.
204
205 21. Retrieve executable code size support for the JIT compiler and fixing
206 some warnings.
207
208 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
209 not work when the shorter character appeared right at the end of the
210 subject string.
211
212 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
213 systems.
214
215 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
216 output it when the /M option is used in pcretest.
217
218 25. The CheckMan script was not being included in the distribution. Also, added
219 an explicit "perl" to run Perl scripts from the PrepareRelease script
220 because this is reportedly needed in Windows.
221
222 26. If study data was being save in a file and studying had not found a set of
223 "starts with" bytes for the pattern, the data written to the file (though
224 never used) was taken from uninitialized memory and so caused valgrind to
225 complain.
226
227 27. Updated RunTest.bat as provided by Sheri Pierce.
228
229 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
230
231 29. Computation of memory usage for the table of capturing group names was
232 giving an unnecessarily large value.
233
234
235 Version 8.20 21-Oct-2011
236 ------------------------
237
238 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
239 a POSIX class. After further experiments with Perl, which convinced me that
240 Perl has bugs and confusions, a closing square bracket is no longer allowed
241 in a POSIX name. This bug also affected patterns with classes that started
242 with full stops.
243
244 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
245 captured substring, but while checking the failing first alternative,
246 substring 1 is temporarily captured. If the output vector supplied to
247 pcre_exec() was not big enough for this capture, the yield of the function
248 was still zero ("insufficient space for captured substrings"). This cannot
249 be totally fixed without adding another stack variable, which seems a lot
250 of expense for a edge case. However, I have improved the situation in cases
251 such as /(a)(b)x|abc/ matched against "abc", where the return code
252 indicates that fewer than the maximum number of slots in the ovector have
253 been set.
254
255 3. Related to (2) above: when there are more back references in a pattern than
256 slots in the output vector, pcre_exec() uses temporary memory during
257 matching, and copies in the captures as far as possible afterwards. It was
258 using the entire output vector, but this conflicts with the specification
259 that only 2/3 is used for passing back captured substrings. Now it uses
260 only the first 2/3, for compatibility. This is, of course, another edge
261 case.
262
263 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
264 main code base, and can be used by building with --enable-jit. When this is
265 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
266 runtime --no-jit option is given.
267
268 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
269 ovector, the return from the function was zero, implying that there were
270 other matches that did not fit. The correct "exactly full" value is now
271 returned.
272
273 6. If a subpattern that was called recursively or as a subroutine contained
274 (*PRUNE) or any other control that caused it to give a non-standard return,
275 invalid errors such as "Error -26 (nested recursion at the same subject
276 position)" or even infinite loops could occur.
277
278 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
279 computing the minimum length on reaching *ACCEPT, and so ended up with the
280 wrong value of 1 rather than 0. Further investigation indicates that
281 computing a minimum subject length in the presence of *ACCEPT is difficult
282 (think back references, subroutine calls), and so I have changed the code
283 so that no minimum is registered for a pattern that contains *ACCEPT.
284
285 8. If (*THEN) was present in the first (true) branch of a conditional group,
286 it was not handled as intended. [But see 16 below.]
287
288 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
289 Sheri Pierce.
290
291 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
292 the first byte in a match must be "a".
293
294 11. Change 17 for 8.13 increased the recursion depth for patterns like
295 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
296 pattern contains any instances of (*THEN). If it does not, the old
297 optimizations are restored. It would be nice to do this on a per-group
298 basis, but at the moment that is not feasible.
299
300 12. In some environments, the output of pcretest -C is CRLF terminated. This
301 broke RunTest's code that checks for the link size. A single white space
302 character after the value is now allowed for.
303
304 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
305 For "fr", it uses the Windows-specific input and output files.
306
307 14. If (*THEN) appeared in a group that was called recursively or as a
308 subroutine, it did not work as intended. [But see next item.]
309
310 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
311 pattern fragments (but not containing any | characters). If A and B are
312 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
313 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
314 D. In other words, Perl considers parentheses that do not contain any |
315 characters to be part of a surrounding alternative, whereas PCRE was
316 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
317 differently. PCRE now behaves in the same way as Perl, except in the case
318 of subroutine/recursion calls such as (?1) which have in any case always
319 been different (but PCRE had them first :-).
320
321 16. Related to 15 above: Perl does not treat the | in a conditional group as
322 creating alternatives. Such a group is treated in the same way as an
323 ordinary group without any | characters when processing (*THEN). PCRE has
324 been changed to match Perl's behaviour.
325
326 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
327 RunGrepTest script failed.
328
329 18. Change 22 for version 13 caused atomic groups to use more stack. This is
330 inevitable for groups that contain captures, but it can lead to a lot of
331 stack use in large patterns. The old behaviour has been restored for atomic
332 groups that do not contain any capturing parentheses.
333
334 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
335 suppress the check for a minimum subject length at run time. (If it was
336 given to pcre_exec() or pcre_dfa_exec() it did work.)
337
338 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
339 fail to work when decoding hex characters in data strings in EBCDIC
340 environments.
341
342 21. It appears that in at least one Mac OS environment, the isxdigit() function
343 is implemented as a macro that evaluates to its argument more than once,
344 contravening the C 90 Standard (I haven't checked a later standard). There
345 was an instance in pcretest which caused it to go wrong when processing
346 \x{...} escapes in subject strings. The has been rewritten to avoid using
347 things like p++ in the argument of isxdigit().
348
349
350 Version 8.13 16-Aug-2011
351 ------------------------
352
353 1. The Unicode data tables have been updated to Unicode 6.0.0.
354
355 2. Two minor typos in pcre_internal.h have been fixed.
356
357 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
358 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
359 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
360
361 4. There were a number of related bugs in the code for matching backrefences
362 caselessly in UTF-8 mode when codes for the characters concerned were
363 different numbers of bytes. For example, U+023A and U+2C65 are an upper
364 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
365 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
366 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
367 2-byte code at the end of the subject (it thought there wasn't enough data
368 left).
369
370 5. Comprehensive information about what went wrong is now returned by
371 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
372 as the output vector has at least 2 elements. The offset of the start of
373 the failing character and a reason code are placed in the vector.
374
375 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
376 now returned is for the first byte of the failing character, instead of the
377 last byte inspected. This is an incompatible change, but I hope it is small
378 enough not to be a problem. It makes the returned offset consistent with
379 pcre_exec() and pcre_dfa_exec().
380
381 7. pcretest now gives a text phrase as well as the error number when
382 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
383 failure, the offset and reason code are output.
384
385 8. When \R was used with a maximizing quantifier it failed to skip backwards
386 over a \r\n pair if the subsequent match failed. Instead, it just skipped
387 back over a single character (\n). This seems wrong (because it treated the
388 two characters as a single entity when going forwards), conflicts with the
389 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
390 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
391 has been changed.
392
393 9. Some internal refactoring has changed the processing so that the handling
394 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
395 time (the PCRE_DOTALL option was changed this way some time ago: version
396 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
397 which was always a bit of a fudge. It also means that there is one less
398 argument for the match() function, which reduces its stack requirements
399 slightly. This change also fixes an incompatibility with Perl: the pattern
400 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
401
402 10. More internal refactoring has drastically reduced the number of recursive
403 calls to match() for possessively repeated groups such as (abc)++ when
404 using pcre_exec().
405
406 11. While implementing 10, a number of bugs in the handling of groups were
407 discovered and fixed:
408
409 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
410 (a|)*(?1) gave a compile-time internal error.
411 ((a|)+)+ did not notice that the outer group could match an empty string.
412 (^a|^)+ was not marked as anchored.
413 (.*a|.*)+ was not marked as matching at start or after a newline.
414
415 12. Yet more internal refactoring has removed another argument from the match()
416 function. Special calls to this function are now indicated by setting a
417 value in a variable in the "match data" data block.
418
419 13. Be more explicit in pcre_study() instead of relying on "default" for
420 opcodes that mean there is no starting character; this means that when new
421 ones are added and accidentally left out of pcre_study(), testing should
422 pick them up.
423
424 14. The -s option of pcretest has been documented for ages as being an old
425 synonym of -m (show memory usage). I have changed it to mean "force study
426 for every regex", that is, assume /S for every regex. This is similar to -i
427 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
428 using it. It makes it easier to run collections of tests with and without
429 study enabled, and thereby test pcre_study() more easily. All the standard
430 tests are now run with and without -s (but some patterns can be marked as
431 "never study" - see 20 below).
432
433 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
434 restoration of the capturing data to the outer values was not happening
435 correctly.
436
437 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
438 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
439 pattern had matched an empty string, and so incorrectly returned a no
440 match.
441
442 17. There was optimizing code for the last branch of non-capturing parentheses,
443 and also for the obeyed branch of a conditional subexpression, which used
444 tail recursion to cut down on stack usage. Unfortunately, now that there is
445 the possibility of (*THEN) occurring in these branches, tail recursion is
446 no longer possible because the return has to be checked for (*THEN). These
447 two optimizations have therefore been removed. [But see 8.20/11 above.]
448
449 18. If a pattern containing \R was studied, it was assumed that \R always
450 matched two bytes, thus causing the minimum subject length to be
451 incorrectly computed because \R can also match just one byte.
452
453 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
454 was incorrectly computed.
455
456 20. If /S is present twice on a test pattern in pcretest input, it now
457 *disables* studying, thereby overriding the use of -s on the command line
458 (see 14 above). This is necessary for one or two tests to keep the output
459 identical in both cases.
460
461 21. When (*ACCEPT) was used in an assertion that matched an empty string and
462 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
463
464 22. When an atomic group that contained a capturing parenthesis was
465 successfully matched, but the branch in which it appeared failed, the
466 capturing was not being forgotten if a higher numbered group was later
467 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
468 group 1 to "a", when in fact it should be unset. This applied to multi-
469 branched capturing and non-capturing groups, repeated or not, and also to
470 positive assertions (capturing in negative assertions does not happen
471 in PCRE) and also to nested atomic groups.
472
473 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
474 subject after a captured substring, to make it easier to tell which of a
475 number of identical substrings has been captured.
476
477 24. The way atomic groups are processed by pcre_exec() has been changed so that
478 if they are repeated, backtracking one repetition now resets captured
479 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
480 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
481 "aaa". Previously, it would have been "a". As part of this code
482 refactoring, the way recursive calls are handled has also been changed.
483
484 25. If an assertion condition captured any substrings, they were not passed
485 back unless some other capturing happened later. For example, if
486 (?(?=(a))a) was matched against "a", no capturing was returned.
487
488 26. When studying a pattern that contained subroutine calls or assertions,
489 the code for finding the minimum length of a possible match was handling
490 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
491 group 1 called group 2 while simultaneously a separate group 2 called group
492 1). A stack overflow occurred in this case. I have fixed this by limiting
493 the recursion depth to 10.
494
495 27. Updated RunTest.bat in the distribution to the version supplied by Tom
496 Fortmann. This supports explicit test numbers on the command line, and has
497 argument validation and error reporting.
498
499 28. An instance of \X with an unlimited repeat could fail if at any point the
500 first character it looked at was a mark character.
501
502 29. Some minor code refactoring concerning Unicode properties and scripts
503 should reduce the stack requirement of match() slightly.
504
505 30. Added the '=' option to pcretest to check the setting of unused capturing
506 slots at the end of the pattern, which are documented as being -1, but are
507 not included in the return count.
508
509 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
510 compiled something random. Now it gives a compile-time error (as does
511 Perl).
512
513 32. A *MARK encountered during the processing of a positive assertion is now
514 recorded and passed back (compatible with Perl).
515
516 33. If --only-matching or --colour was set on a pcregrep call whose pattern
517 had alternative anchored branches, the search for a second match in a line
518 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
519 matched the line "0102" twice. The same bug affected patterns that started
520 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
521 twice.
522
523 34. Previously, PCRE did not allow quantification of assertions. However, Perl
524 does, and because of capturing effects, quantifying parenthesized
525 assertions may at times be useful. Quantifiers are now allowed for
526 parenthesized assertions.
527
528 35. A minor code tidy in pcre_compile() when checking options for \R usage.
529
530 36. \g was being checked for fancy things in a character class, when it should
531 just be a literal "g".
532
533 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
534 appearance of a nested POSIX class supersedes an apparent external class.
535 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
536 unescaped square brackets may also appear as part of class names. For
537 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
538 more like Perl. (But see 8.20/1 above.)
539
540 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
541 was because it thought it was \N{name}, which is not supported).
542
543 39. Add minix to OS list not supporting the -S option in pcretest.
544
545 40. PCRE tries to detect cases of infinite recursion at compile time, but it
546 cannot analyze patterns in sufficient detail to catch mutual recursions
547 such as ((?1))((?2)). There is now a runtime test that gives an error if a
548 subgroup is called recursively as a subpattern for a second time at the
549 same position in the subject string. In previous releases this might have
550 been caught by the recursion limit, or it might have run out of stack.
551
552 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
553 happen only once. PCRE was, however incorrectly giving a compile time error
554 "recursive call could loop indefinitely" because it cannot analyze the
555 pattern in sufficient detail. The compile time test no longer happens when
556 PCRE is compiling a conditional subpattern, but actual runaway loops are
557 now caught at runtime (see 40 above).
558
559 42. It seems that Perl allows any characters other than a closing parenthesis
560 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
561 has been changed to be the same.
562
563 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
564 as not to get warnings when autogen.sh is called. Also changed
565 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
566
567 44. To help people who use pcregrep to scan files containing exceedingly long
568 lines, the following changes have been made:
569
570 (a) The default value of the buffer size parameter has been increased from
571 8K to 20K. (The actual buffer used is three times this size.)
572
573 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
574 PCRE is built.
575
576 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
577 to be set at run time.
578
579 (d) Numerical values in pcregrep options can be followed by K or M, for
580 example --buffer-size=50K.
581
582 (e) If a line being scanned overflows pcregrep's buffer, an error is now
583 given and the return code is set to 2.
584
585 45. Add a pointer to the latest mark to the callout data block.
586
587 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
588 partial match of an empty string instead of no match. This was specific to
589 the use of ".".
590
591 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
592 complete match instead of a partial match. This bug was dependent on both
593 the PCRE_UTF8 and PCRE_DOTALL options being set.
594
595 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
596 starting byte set, because \b was not being ignored.
597
598
599 Version 8.12 15-Jan-2011
600 ------------------------
601
602 1. Fixed some typos in the markup of the man pages, and wrote a script that
603 checks for such things as part of the documentation building process.
604
605 2. On a big-endian 64-bit system, pcregrep did not correctly process the
606 --match-limit and --recursion-limit options (added for 8.11). In
607 particular, this made one of the standard tests fail. (The integer value
608 went into the wrong half of a long int.)
609
610 3. If the --colour option was given to pcregrep with -v (invert match), it
611 did strange things, either producing crazy output, or crashing. It should,
612 of course, ignore a request for colour when reporting lines that do not
613 match.
614
615 4. Another pcregrep bug caused similar problems if --colour was specified with
616 -M (multiline) and the pattern match finished with a line ending.
617
618 5. In pcregrep, when a pattern that ended with a literal newline sequence was
619 matched in multiline mode, the following line was shown as part of the
620 match. This seems wrong, so I have changed it.
621
622 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
623 the check for further matches in the same line (so they could be coloured)
624 to overrun the end of the current line. If another match was found, it was
625 incorrectly shown (and then shown again when found in the next line).
626
627 7. If pcregrep was compiled under Windows, there was a reference to the
628 function pcregrep_exit() before it was defined. I am assuming this was
629 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
630 reported by a user. I've moved the definition above the reference.
631
632
633 Version 8.11 10-Dec-2010
634 ------------------------
635
636 1. (*THEN) was not working properly if there were untried alternatives prior
637 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
638 backtracked to try for "b" instead of moving to the next alternative branch
639 at the same level (in this case, to look for "c"). The Perl documentation
640 is clear that when (*THEN) is backtracked onto, it goes to the "next
641 alternative in the innermost enclosing group".
642
643 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
644 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
645 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
646 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
647 (*THEN).
648
649 3. If \s appeared in a character class, it removed the VT character from
650 the class, even if it had been included by some previous item, for example
651 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
652 of \s, but is part of the POSIX "space" class.)
653
654 4. A partial match never returns an empty string (because you can always
655 match an empty string at the end of the subject); however the checking for
656 an empty string was starting at the "start of match" point. This has been
657 changed to the "earliest inspected character" point, because the returned
658 data for a partial match starts at this character. This means that, for
659 example, /(?<=abc)def/ gives a partial match for the subject "abc"
660 (previously it gave "no match").
661
662 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
663 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
664 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
665 has an implication that the given string is incomplete (because a partial
666 match is preferred over a full match). For this reason, these items now
667 give a partial match in this situation. [Aside: previously, the one case
668 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
669 match rather than a full match, which was wrong by the old rules, but is
670 now correct.]
671
672 6. There was a bug in the handling of #-introduced comments, recognized when
673 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
674 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
675 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
676 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
677 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
678 places in pcre_compile().
679
680 7. Related to (6) above, when pcre_compile() was skipping #-introduced
681 comments when looking ahead for named forward references to subpatterns,
682 the only newline sequence it recognized was NL. It now handles newlines
683 according to the set newline convention.
684
685 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
686 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
687 cater for a lack of strerror(). These oversights have been fixed.
688
689 9. Added --match-limit and --recursion-limit to pcregrep.
690
691 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
692
693 11. When the -o option was used, pcregrep was setting a return code of 1, even
694 when matches were found, and --line-buffered was not being honoured.
695
696 12. Added an optional parentheses number to the -o and --only-matching options
697 of pcregrep.
698
699 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
700 can match an empty string. The code to do it in pcretest and pcredemo
701 needed fixing:
702
703 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
704 only one byte after an empty string match just before CRLF (this case
705 just got forgotten; "any" and "anycrlf" were OK).
706
707 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
708 mode when an empty string match preceded an ASCII character followed by
709 a non-ASCII character. (The code for advancing by one character rather
710 than one byte was nonsense.)
711
712 (c) The pcredemo.c sample program did not have any code at all to handle
713 the cases when CRLF is a valid newline sequence.
714
715 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
716 as a starting offset was within the subject string. There is now a new
717 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
718 negative or greater than the length of the string. In order to test this,
719 pcretest is extended to allow the setting of negative starting offsets.
720
721 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
722 starting offset points to the beginning of a UTF-8 character was
723 unnecessarily clumsy. I tidied it up.
724
725 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
726 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
727
728 17. Nobody had reported that the --include_dir option, which was added in
729 release 7.7 should have been called --include-dir (hyphen, not underscore)
730 for compatibility with GNU grep. I have changed it to --include-dir, but
731 left --include_dir as an undocumented synonym, and the same for
732 --exclude-dir, though that is not available in GNU grep, at least as of
733 release 2.5.4.
734
735 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
736 characters from a string of bytes) have been redefined so as not to use
737 loops, in order to improve performance in some environments. At the same
738 time, I abstracted some of the common code into auxiliary macros to save
739 repetition (this should not affect the compiled code).
740
741 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
742 compile-time error is now given if \c is not followed by an ASCII
743 character, that is, a byte less than 128. (In EBCDIC mode, the code is
744 different, and any byte value is allowed.)
745
746 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
747 START_OPTIMIZE option, which is now allowed at compile time - but just
748 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
749 to pcregrep and other applications that have no direct access to PCRE
750 options. The new /Y option in pcretest sets this option when calling
751 pcre_compile().
752
753 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
754 back references. Groups containing recursive back references were forced to
755 be atomic by that change, but in the case of named groups, the amount of
756 memory required was incorrectly computed, leading to "Failed: internal
757 error: code overflow". This has been fixed.
758
759 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
760 pcretest.c, to avoid build problems in some Borland environments.
761
762
763 Version 8.10 25-Jun-2010
764 ------------------------
765
766 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
767 THEN.
768
769 2. (*ACCEPT) was not working when inside an atomic group.
770
771 3. Inside a character class, \B is treated as a literal by default, but
772 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
773 causes the error). The code is unchanged, but I tidied the documentation.
774
775 4. Inside a character class, PCRE always treated \R and \X as literals,
776 whereas Perl faults them if its -w option is set. I have changed PCRE so
777 that it faults them when PCRE_EXTRA is set.
778
779 5. Added support for \N, which always matches any character other than
780 newline. (It is the same as "." when PCRE_DOTALL is not set.)
781
782 6. When compiling pcregrep with newer versions of gcc which may have
783 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
784 declared with attribute warn_unused_result" were given. Just casting the
785 result to (void) does not stop the warnings; a more elaborate fudge is
786 needed. I've used a macro to implement this.
787
788 7. Minor change to pcretest.c to avoid a compiler warning.
789
790 8. Added four artifical Unicode properties to help with an option to make
791 \s etc use properties (see next item). The new properties are: Xan
792 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
793
794 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
795 use Unicode properties. (*UCP) at the start of a pattern can be used to set
796 this option. Modified pcretest to add /W to test this facility. Added
797 REG_UCP to make it available via the POSIX interface.
798
799 10. Added --line-buffered to pcregrep.
800
801 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
802 studied, and the match started with a letter with a code point greater than
803 127 whose first byte was different to the first byte of the other case of
804 the letter, the other case of this starting letter was not recognized
805 (#976).
806
807 12. If a pattern that was studied started with a repeated Unicode property
808 test, for example, \p{Nd}+, there was the theoretical possibility of
809 setting up an incorrect bitmap of starting bytes, but fortunately it could
810 not have actually happened in practice until change 8 above was made (it
811 added property types that matched character-matching opcodes).
812
813 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
814 possible starting bytes for non-anchored patterns.
815
816 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
817 \R, and also a number of cases that involve Unicode properties, both
818 explicit and implicit when PCRE_UCP is set.
819
820 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
821 input, it could crash or give wrong results if characters with values
822 greater than 0xc0 were present in the subject string. (Detail: it assumed
823 UTF-8 input when processing these items.)
824
825 16. Added a lot of (int) casts to avoid compiler warnings in systems where
826 size_t is 64-bit (#991).
827
828 17. Added a check for running out of memory when PCRE is compiled with
829 --disable-stack-for-recursion (#990).
830
831 18. If the last data line in a file for pcretest does not have a newline on
832 the end, a newline was missing in the output.
833
834 19. The default pcre_chartables.c file recognizes only ASCII characters (values
835 less than 128) in its various bitmaps. However, there is a facility for
836 generating tables according to the current locale when PCRE is compiled. It
837 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
838 characters, are recognized by isspace() and therefore were getting set in
839 these tables, and indeed these tables seem to approximate to ISO 8859. This
840 caused a problem in UTF-8 mode when pcre_study() was used to create a list
841 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
842 which of course cannot start UTF-8 characters. I have changed the code so
843 that only real ASCII characters (less than 128) and the correct starting
844 bytes for UTF-8 encodings are set for characters greater than 127 when in
845 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
846 altogether.)
847
848 20. Added the /T option to pcretest so as to be able to run tests with non-
849 standard character tables, thus making it possible to include the tests
850 used for 19 above in the standard set of tests.
851
852 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
853 reference to a subpattern the other side of a comment that contains an
854 opening parenthesis caused either an internal compiling error, or a
855 reference to the wrong subpattern.
856
857
858 Version 8.02 19-Mar-2010
859 ------------------------
860
861 1. The Unicode data tables have been updated to Unicode 5.2.0.
862
863 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
864 configured.
865
866 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
867 original author of that file, following a query about its status.
868
869 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
870 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
871
872 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
873 quantifier applied to a forward-referencing subroutine call, could compile
874 incorrect code or give the error "internal error: previously-checked
875 referenced subpattern not found".
876
877 6. Both MS Visual Studio and Symbian OS have problems with initializing
878 variables to point to external functions. For these systems, therefore,
879 pcre_malloc etc. are now initialized to local functions that call the
880 relevant global functions.
881
882 7. There were two entries missing in the vectors called coptable and poptable
883 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
884 I've fixed the data, and added a kludgy way of testing at compile time that
885 the lengths are correct (equal to the number of opcodes).
886
887 8. Following on from 7, I added a similar kludge to check the length of the
888 eint vector in pcreposix.c.
889
890 9. Error texts for pcre_compile() are held as one long string to avoid too
891 much relocation at load time. To find a text, the string is searched,
892 counting zeros. There was no check for running off the end of the string,
893 which could happen if a new error number was added without updating the
894 string.
895
896 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
897
898 11. \K was not working if it appeared in an atomic group or in a group that
899 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
900 \K is "not well defined" if used in an assertion. PCRE now accepts it if
901 the assertion is positive, but not if it is negative.
902
903 12. Change 11 fortuitously reduced the size of the stack frame used in the
904 "match()" function of pcre_exec.c by one pointer. Forthcoming
905 implementation of support for (*MARK) will need an extra pointer on the
906 stack; I have reserved it now, so that the stack frame size does not
907 decrease.
908
909 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
910 item in branch that calls a recursion is a subroutine call - as in the
911 second branch in the above example - was incorrectly given the compile-
912 time error "recursive call could loop indefinitely" because pcre_compile()
913 was not correctly checking the subroutine for matching a non-empty string.
914
915 14. The checks for overrunning compiling workspace could trigger after an
916 overrun had occurred. This is a "should never occur" error, but it can be
917 triggered by pathological patterns such as hundreds of nested parentheses.
918 The checks now trigger 100 bytes before the end of the workspace.
919
920 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
921
922
923 Version 8.01 19-Jan-2010
924 ------------------------
925
926 1. If a pattern contained a conditional subpattern with only one branch (in
927 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
928 computed the wrong minimum data length (which is of course zero for such
929 subpatterns). This could cause incorrect "no match" results.
930
931 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
932 the pattern is reset in the first branch, pcre_compile() failed with
933 "internal error: code overflow at offset...". This happened only when
934 the reset was to the original external option setting. (An optimization
935 abstracts leading options settings into an external setting, which was the
936 cause of this.)
937
938 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
939 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
940 assertion pattern did not match (meaning that the assertion was true), it
941 was incorrectly treated as false if the SKIP had been reached during the
942 matching. This also applied to assertions used as conditions.
943
944 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
945 assertion subpattern, including such a pattern used as a condition,
946 unpredictable results occurred, instead of the error return
947 PCRE_ERROR_DFA_UITEM.
948
949 5. The C++ GlobalReplace function was not working like Perl for the special
950 situation when an empty string is matched. It now does the fancy magic
951 stuff that is necessary.
952
953 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
954 removed. (These were left over from very, very early versions of PCRE.)
955
956 7. Some cosmetic changes to the code to make life easier when compiling it
957 as part of something else:
958
959 (a) Change DEBUG to PCRE_DEBUG.
960
961 (b) In pcre_compile(), rename the member of the "branch_chain" structure
962 called "current" as "current_branch", to prevent a collision with the
963 Linux macro when compiled as a kernel module.
964
965 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
966 prevent a collision with the Linux macro when compiled as a kernel
967 module.
968
969 8. In pcre_compile() there are some checks for integer overflows that used to
970 cast potentially large values to (double). This has been changed to that
971 when building, a check for int64_t is made, and if it is found, it is used
972 instead, thus avoiding the use of floating point arithmetic. (There is no
973 other use of FP in PCRE.) If int64_t is not found, the fallback is to
974 double.
975
976 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
977 2005 (difference between two addresses compared to an unsigned value).
978
979 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
980 custom one, because of the following reported problem in Windows:
981
982 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
983 under Win32.
984 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
985 therefore missing the function definition.
986 - The compiler thus generates a "C" signature for the test function.
987 - The linker fails to find the "C" function.
988 - PCRE fails to configure if asked to do so against libbz2.
989
990 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
991 messages were output:
992
993 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
994 rerunning libtoolize, to keep the correct libtool macros in-tree.
995 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
996
997 I have done both of these things.
998
999 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1000 most of the time, it *can* run out if it is given a pattern that contains a
1001 runaway infinite recursion. I updated the discussion in the pcrestack man
1002 page.
1003
1004 13. Now that we have gone to the x.xx style of version numbers, the minor
1005 version may start with zero. Using 08 or 09 is a bad idea because users
1006 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1007 interpreted as invalid octal numbers. I've updated the previous comment in
1008 configure.ac, and also added a check that gives an error if 08 or 09 are
1009 used.
1010
1011 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1012 causing partial matching to fail when the end of the subject matched \W
1013 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1014
1015 15. There were some discrepancies between the declarations in pcre_internal.h
1016 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1017 their definitions. The declarations used "const uschar *" and the
1018 definitions used USPTR. Even though USPTR is normally defined as "const
1019 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1020 reported that: "This difference in casting confuses some C++ compilers, for
1021 example, SunCC recognizes above declarations as different functions and
1022 generates broken code for hbpcre." I have changed the declarations to use
1023 USPTR.
1024
1025 16. GNU libtool is named differently on some systems. The autogen.sh script now
1026 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1027 (FreeBSD).
1028
1029 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1030 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1031 comment: "Figure out how to create a longlong from a string: strtoll and
1032 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1033 instance, but it only takes 2 args instead of 3!"
1034
1035 18. A subtle bug concerned with back references has been fixed by a change of
1036 specification, with a corresponding code fix. A pattern such as
1037 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1038 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1039 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1040 same bug. Such groups have to be quantified to be useful, or contained
1041 inside another quantified group. (If there's no repetition, the reference
1042 can never match.) The problem arises because, having left the group and
1043 moved on to the rest of the pattern, a later failure that backtracks into
1044 the group uses the captured value from the final iteration of the group
1045 rather than the correct earlier one. I have fixed this in PCRE by forcing
1046 any group that contains a reference to itself to be an atomic group; that
1047 is, there cannot be any backtracking into it once it has completed. This is
1048 similar to recursive and subroutine calls.
1049
1050
1051 Version 8.00 19-Oct-09
1052 ----------------------
1053
1054 1. The table for translating pcre_compile() error codes into POSIX error codes
1055 was out-of-date, and there was no check on the pcre_compile() error code
1056 being within the table. This could lead to an OK return being given in
1057 error.
1058
1059 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1060 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1061 in a Windows environment.
1062
1063 3. The pcregrep --count option prints the count for each file even when it is
1064 zero, as does GNU grep. However, pcregrep was also printing all files when
1065 --files-with-matches was added. Now, when both options are given, it prints
1066 counts only for those files that have at least one match. (GNU grep just
1067 prints the file name in this circumstance, but including the count seems
1068 more useful - otherwise, why use --count?) Also ensured that the
1069 combination -clh just lists non-zero counts, with no names.
1070
1071 4. The long form of the pcregrep -F option was incorrectly implemented as
1072 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1073 but it seems right to fix it, and I didn't think it was worth preserving
1074 the old behaviour.
1075
1076 5. The command line items --regex=pattern and --regexp=pattern were not
1077 recognized by pcregrep, which required --regex pattern or --regexp pattern
1078 (with a space rather than an '='). The man page documented the '=' forms,
1079 which are compatible with GNU grep; these now work.
1080
1081 6. No libpcreposix.pc file was created for pkg-config; there was just
1082 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1083
1084 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1085 when UCP support is not needed, by modifying the Python script that
1086 generates it from Unicode data files. This should not matter if the module
1087 is correctly used as a library, but I received one complaint about 50K of
1088 unwanted data. My guess is that the person linked everything into his
1089 program rather than using a library. Anyway, it does no harm.
1090
1091 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1092 was a minimum greater than 1 for a wide character in a possessive
1093 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1094 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1095 character. Chaos in the form of incorrect output or a compiling loop could
1096 result.
1097
1098 9. The restrictions on what a pattern can contain when partial matching is
1099 requested for pcre_exec() have been removed. All patterns can now be
1100 partially matched by this function. In addition, if there are at least two
1101 slots in the offset vector, the offset of the earliest inspected character
1102 for the match and the offset of the end of the subject are set in them when
1103 PCRE_ERROR_PARTIAL is returned.
1104
1105 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1106 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1107 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1108 and may be more useful for multi-segment matching.
1109
1110 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1111 used to be given if ever the end of the subject was reached; now it is
1112 given only if matching could not proceed because another character was
1113 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1114 string "Z", which now yields "no match" instead of "partial match". In the
1115 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1116 final character ended with (*FAIL).
1117
1118 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1119 if the pattern had a "must contain" character that was already found in the
1120 earlier partial match, unless partial matching was again requested. For
1121 example, with the pattern /dog.(body)?/, the "must contain" character is
1122 "g". If the first part-match was for the string "dog", restarting with
1123 "sbody" failed. This bug has been fixed.
1124
1125 13. The string returned by pcre_dfa_exec() after a partial match has been
1126 changed so that it starts at the first inspected character rather than the
1127 first character of the match. This makes a difference only if the pattern
1128 starts with a lookbehind assertion or \b or \B (\K is not supported by
1129 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1130 matching functions compatible, and I think it's the right thing to do.
1131
1132 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1133 so that the demonstration program is easily available in environments where
1134 PCRE has not been installed from source.
1135
1136 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1137 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1138 library.
1139
1140 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1141 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1142 is not the first non-POSIX option to be added. Clearly some people find
1143 these options useful.
1144
1145 17. If a caller to the POSIX matching function regexec() passes a non-zero
1146 value for nmatch with a NULL value for pmatch, the value of
1147 nmatch is forced to zero.
1148
1149 18. RunGrepTest did not have a test for the availability of the -u option of
1150 the diff command, as RunTest does. It now checks in the same way as
1151 RunTest, and also checks for the -b option.
1152
1153 19. If an odd number of negated classes containing just a single character
1154 interposed, within parentheses, between a forward reference to a named
1155 subpattern and the definition of the subpattern, compilation crashed with
1156 an internal error, complaining that it could not find the referenced
1157 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1158 [The bug was that it was starting one character too far in when skipping
1159 over the character class, thus treating the ] as data rather than
1160 terminating the class. This meant it could skip too much.]
1161
1162 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1163 /g option in pcretest when the pattern contains \K, which makes it possible
1164 to have an empty string match not at the start, even when the pattern is
1165 anchored. Updated pcretest and pcredemo to use this option.
1166
1167 21. If the maximum number of capturing subpatterns in a recursion was greater
1168 than the maximum at the outer level, the higher number was returned, but
1169 with unset values at the outer level. The correct (outer level) value is
1170 now given.
1171
1172 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1173 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1174 make it do so. The string so far is captured, making this feature
1175 compatible with Perl.
1176
1177 23. The tests have been re-organized, adding tests 11 and 12, to make it
1178 possible to check the Perl 5.10 features against Perl 5.10.
1179
1180 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1181 pattern matches a fixed length string. PCRE did not allow this; now it
1182 does. Neither allows recursion.
1183
1184 25. I finally figured out how to implement a request to provide the minimum
1185 length of subject string that was needed in order to match a given pattern.
1186 (It was back references and recursion that I had previously got hung up
1187 on.) This code has now been added to pcre_study(); it finds a lower bound
1188 to the length of subject needed. It is not necessarily the greatest lower
1189 bound, but using it to avoid searching strings that are too short does give
1190 some useful speed-ups. The value is available to calling programs via
1191 pcre_fullinfo().
1192
1193 26. While implementing 25, I discovered to my embarrassment that pcretest had
1194 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1195 study optimizations had never been tested with that matching function.
1196 Oops. What is worse, even when it was passed study data, there was a bug in
1197 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1198 were also very few tests of studied patterns with pcre_dfa_exec().
1199
1200 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1201 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1202 on the other side of the coin, they are no longer allowed to have different
1203 names, because these cannot be distinguished in PCRE, and this has caused
1204 confusion. (This is a difference from Perl.)
1205
1206 28. When duplicate subpattern names are present (necessarily with different
1207 numbers, as required by 27 above), and a test is made by name in a
1208 conditional pattern, either for a subpattern having been matched, or for
1209 recursion in such a pattern, all the associated numbered subpatterns are
1210 tested, and the overall condition is true if the condition is true for any
1211 one of them. This is the way Perl works, and is also more like the way
1212 testing by number works.
1213
1214
1215 Version 7.9 11-Apr-09
1216 ---------------------
1217
1218 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1219 (pcretest), all targets were linked against these libraries. This included
1220 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1221 libraries. This caused unwanted dependencies to be created. This problem
1222 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1223 pcretest is linked with readline.
1224
1225 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1226 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1227 moved outside it again, because FALSE and TRUE are already defined in AIX,
1228 but BOOL is not.
1229
1230 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1231 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1232
1233 4. The pcregrep documentation said spaces were inserted as well as colons (or
1234 hyphens) following file names and line numbers when outputting matching
1235 lines. This is not true; no spaces are inserted. I have also clarified the
1236 wording for the --colour (or --color) option.
1237
1238 5. In pcregrep, when --colour was used with -o, the list of matching strings
1239 was not coloured; this is different to GNU grep, so I have changed it to be
1240 the same.
1241
1242 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1243 each matching line was coloured. Now it goes on to look for further matches
1244 of any of the test patterns, which is the same behaviour as GNU grep.
1245
1246 7. A pattern that could match an empty string could cause pcregrep to loop; it
1247 doesn't make sense to accept an empty string match in pcregrep, so I have
1248 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1249 seems to be how GNU grep behaves.
1250
1251 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1252 start or after a newline", because the conditional assertion was not being
1253 correctly handled. The rule now is that both the assertion and what follows
1254 in the first alternative must satisfy the test.
1255
1256 9. If auto-callout was enabled in a pattern with a conditional group whose
1257 condition was an assertion, PCRE could crash during matching, both with
1258 pcre_exec() and pcre_dfa_exec().
1259
1260 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1261 used for matching.
1262
1263 11. Unicode property support in character classes was not working for
1264 characters (bytes) greater than 127 when not in UTF-8 mode.
1265
1266 12. Added the -M command line option to pcretest.
1267
1268 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1269
1270 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1271
1272 16. Added comments and documentation about mis-use of no_arg in the C++
1273 wrapper.
1274
1275 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1276 from Martin Jerabek that uses macro names for all relevant character and
1277 string constants.
1278
1279 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1280 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1281 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1282 these, but not everybody uses configure.
1283
1284 19. A conditional group that had only one branch was not being correctly
1285 recognized as an item that could match an empty string. This meant that an
1286 enclosing group might also not be so recognized, causing infinite looping
1287 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1288 with the subject "ab", where knowledge that the repeated group can match
1289 nothing is needed in order to break the loop.
1290
1291 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1292 exec(), but without supplying a callout function, matching went wrong.
1293
1294 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1295 leak if the size of the offset vector was greater than 30. When the vector
1296 is smaller, the saved offsets during recursion go onto a local stack
1297 vector, but for larger vectors malloc() is used. It was failing to free
1298 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1299 error, in fact).
1300
1301 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1302 heapframe that is used only when UTF-8 support is enabled. This caused no
1303 problem, but was untidy.
1304
1305 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1306 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1307 included within another project.
1308
1309 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1310 slightly modified by me:
1311
1312 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1313 not building pcregrep.
1314
1315 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1316 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1317
1318 25. Forward references, both numeric and by name, in patterns that made use of
1319 duplicate group numbers, could behave incorrectly or give incorrect errors,
1320 because when scanning forward to find the reference group, PCRE was not
1321 taking into account the duplicate group numbers. A pattern such as
1322 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1323
1324 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1325 the feature of a custom pointer more persuasive (as requested by a user).
1326
1327 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1328 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1329 pre-defined.
1330
1331 28. Added support for (*UTF8) at the start of a pattern.
1332
1333 29. Arrange for flags added by the "release type" setting in CMake to be shown
1334 in the configuration summary.
1335
1336
1337 Version 7.8 05-Sep-08
1338 ---------------------
1339
1340 1. Replaced UCP searching code with optimized version as implemented for Ad
1341 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1342 stage table and inline lookup instead of a function, giving speed ups of 2
1343 to 5 times on some simple patterns that I tested. Permission was given to
1344 distribute the MultiStage2.py script that generates the tables (it's not in
1345 the tarball, but is in the Subversion repository).
1346
1347 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1348 scripts.
1349
1350 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1351 a group with a zero qualifier. The result of the study could be incorrect,
1352 or the function might crash, depending on the pattern.
1353
1354 4. Caseless matching was not working for non-ASCII characters in back
1355 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1356 It now works when Unicode Property Support is available.
1357
1358 5. In pcretest, an escape such as \x{de} in the data was always generating
1359 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1360 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1361 truncation.
1362
1363 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1364
1365 7. Added two (int) casts to pcregrep when printing the difference of two
1366 pointers, in case they are 64-bit values.
1367
1368 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1369 test 2 if it fails.
1370
1371 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1372 and a #define of that name to empty if it is not externally set. This is to
1373 allow users of MSVC to set it if necessary.
1374
1375 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1376 the convenience functions in the pcre_get.c source file.
1377
1378 11. An option change at the start of a pattern that had top-level alternatives
1379 could cause overwriting and/or a crash. This command provoked a crash in
1380 some environments:
1381
1382 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1383
1384 This potential security problem was recorded as CVE-2008-2371.
1385
1386 12. For a pattern where the match had to start at the beginning or immediately
1387 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1388 pcre_dfa_exec() could read past the end of the passed subject if there was
1389 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1390 pcretest so that it places the subject at the end of its malloc-ed buffer.
1391
1392 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1393 exec() might read past the end of the data buffer in UTF-8 mode.
1394
1395 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1396 the data contained the byte 0x85 as part of a UTF-8 character within its
1397 first line. This applied both to normal and DFA matching.
1398
1399 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1400 /^[^d]*?$/8 failed to match "abc".
1401
1402 16. Added a missing copyright notice to pcrecpp_internal.h.
1403
1404 17. Make it more clear in the documentation that values returned from
1405 pcre_exec() in ovector are byte offsets, not character counts.
1406
1407 18. Tidied a few places to stop certain compilers from issuing warnings.
1408
1409 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1410 supplied by Stefan Weber. I made a further small update for 7.8 because
1411 there is a change of source arrangements: the pcre_searchfuncs.c module is
1412 replaced by pcre_ucd.c.
1413
1414
1415 Version 7.7 07-May-08
1416 ---------------------
1417
1418 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1419 a string to a long long, pretend we don't even have a long long." This is
1420 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1421
1422 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1423 pre-7.6 versions, which defined a global no_arg variable instead of putting
1424 it in the RE class. (See also #8 below.)
1425
1426 3. Remove a line of dead code, identified by coverity and reported by Nuno
1427 Lopes.
1428
1429 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1430
1431 (1) The include/exclude patterns were being applied to the whole pathnames
1432 of files, instead of just to the final components.
1433
1434 (2) If there was more than one level of directory, the subdirectories were
1435 skipped unless they satisfied the include/exclude conditions. This is
1436 inconsistent with GNU grep (and could even be seen as contrary to the
1437 pcregrep specification - which I improved to make it absolutely clear).
1438 The action now is always to scan all levels of directory, and just
1439 apply the include/exclude patterns to regular files.
1440
1441 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1442 --exclude_dir in the tests to avoid scanning .svn directories.
1443
1444 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1445 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1446 doesn't support NULs in patterns.
1447
1448 7. Added some missing "const"s to declarations of static tables in
1449 pcre_compile.c and pcre_dfa_exec.c.
1450
1451 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1452 caused by fix #2 above. (Subsequently also a second patch to fix the
1453 first patch. And a third patch - this was a messy problem.)
1454
1455 9. Applied Craig's patch to remove the use of push_back().
1456
1457 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1458 matching function regexec().
1459
1460 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1461 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1462 references. PCRE supports relative numbers with this syntax (I don't think
1463 Oniguruma does).
1464
1465 12. Previously, a group with a zero repeat such as (...){0} was completely
1466 omitted from the compiled regex. However, this means that if the group
1467 was called as a subroutine from elsewhere in the pattern, things went wrong
1468 (an internal error was given). Such groups are now left in the compiled
1469 pattern, with a new opcode that causes them to be skipped at execution
1470 time.
1471
1472 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1473 to the way PCRE behaves:
1474
1475 (a) A lone ] character is dis-allowed (Perl treats it as data).
1476
1477 (b) A back reference to an unmatched subpattern matches an empty string
1478 (Perl fails the current match path).
1479
1480 (c) A data ] in a character class must be notated as \] because if the
1481 first data character in a class is ], it defines an empty class. (In
1482 Perl it is not possible to have an empty class.) The empty class []
1483 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1484 The negative empty class [^] matches any one character, independently
1485 of the DOTALL setting.
1486
1487 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1488 non-existent subpattern following a character class starting with ']' and
1489 containing () gave an internal compiling error instead of "reference to
1490 non-existent subpattern". Fortunately, when the pattern did exist, the
1491 compiled code was correct. (When scanning forwards to check for the
1492 existencd of the subpattern, it was treating the data ']' as terminating
1493 the class, so got the count wrong. When actually compiling, the reference
1494 was subsequently set up correctly.)
1495
1496 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1497 it was being rejected as not supported by pcre_dfa_exec(), even though
1498 other assertions are supported. I have made pcre_dfa_exec() support
1499 (*FAIL).
1500
1501 16. The implementation of 13c above involved the invention of a new opcode,
1502 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1503 cannot be changed at match time, I realized I could make a small
1504 improvement to matching performance by compiling OP_ALLANY instead of
1505 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1506 on the OP_ANY path.
1507
1508 17. Compiling pcretest on Windows with readline support failed without the
1509 following two fixes: (1) Make the unistd.h include conditional on
1510 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1511
1512 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1513 ncurses library to be included for pcretest when ReadLine support is
1514 requested, but also to allow for it to be overridden. This patch came from
1515 Daniel Bergström.
1516
1517 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1518 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1519 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1520 spotting this.
1521
1522
1523 Version 7.6 28-Jan-08
1524 ---------------------
1525
1526 1. A character class containing a very large number of characters with
1527 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1528 overflow.
1529
1530 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1531 HAVE_LONG_LONG is not defined.
1532
1533 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1534 bring them up to date and include new features. This patch includes:
1535
1536 - Fixed PH's badly added libz and libbz2 support.
1537 - Fixed a problem with static linking.
1538 - Added pcredemo. [But later removed - see 7 below.]
1539 - Fixed dftables problem and added an option.
1540 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1541 HAVE_LONG_LONG.
1542 - Added readline support for pcretest.
1543 - Added an listing of the option settings after cmake has run.
1544
1545 4. A user submitted a patch to Makefile that makes it easy to create
1546 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1547 Makefile.am that cause it to include this special target, without
1548 affecting anything else. Note that the same mingw target plus all
1549 the other distribution libraries and programs are now supported
1550 when configuring with CMake (see 6 below) instead of with
1551 Configure/Make.
1552
1553 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1554 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1555 exported in the Windows port". It has not yet been confirmed that the patch
1556 solves the problem, but it does no harm.
1557
1558 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1559 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1560 with CMake, and also correct the comment about stack recursion.
1561
1562 7. Remove the automatic building of pcredemo from the ./configure system and
1563 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1564 of a program that users should build themselves after PCRE is installed, so
1565 building it automatically is not really right. What is more, it gave
1566 trouble in some build environments.
1567
1568 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1569
1570
1571 Version 7.5 10-Jan-08
1572 ---------------------
1573
1574 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1575 values in parens when parsing an RE using the C++ wrapper."
1576
1577 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1578 Characters greater than 255 were excluded from the class instead of being
1579 included.
1580
1581 3. The same bug as (2) above applied to negated POSIX classes such as
1582 [:^space:].
1583
1584 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1585 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1586 I have changed it.
1587
1588 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1589 first named subpattern) and a construct such as (?&a) would reference the
1590 first named subpattern whose name started with "a" (in other words, the
1591 length check was missing). Both these problems are fixed. "Subpattern name
1592 expected" is now given for (?&) (a zero-length name), and this patch also
1593 makes it give the same error for \k'' (previously it complained that that
1594 was a reference to a non-existent subpattern).
1595
1596 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1597 this is right because (?- can be followed by option settings as well as by
1598 digits. I have, however, made the messages clearer.
1599
1600 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1601 than the number used in the conditional) now cause a compile-time error.
1602 This is actually not compatible with Perl, which accepts such patterns, but
1603 treats the conditional as always being FALSE (as PCRE used to), but it
1604 seems to me that giving a diagnostic is better.
1605
1606 8. Change "alphameric" to the more common word "alphanumeric" in comments
1607 and messages.
1608
1609 9. Fix two occurrences of "backslash" in comments that should have been
1610 "backspace".
1611
1612 10. Remove two redundant lines of code that can never be obeyed (their function
1613 was moved elsewhere).
1614
1615 11. The program that makes PCRE's Unicode character property table had a bug
1616 which caused it to generate incorrect table entries for sequences of
1617 characters that have the same character type, but are in different scripts.
1618 It amalgamated them into a single range, with the script of the first of
1619 them. In other words, some characters were in the wrong script. There were
1620 thirteen such cases, affecting characters in the following ranges:
1621
1622 U+002b0 - U+002c1
1623 U+0060c - U+0060d
1624 U+0061e - U+00612
1625 U+0064b - U+0065e
1626 U+0074d - U+0076d
1627 U+01800 - U+01805
1628 U+01d00 - U+01d77
1629 U+01d9b - U+01dbf
1630 U+0200b - U+0200f
1631 U+030fc - U+030fe
1632 U+03260 - U+0327f
1633 U+0fb46 - U+0fbb1
1634 U+10450 - U+1049d
1635
1636 12. The -o option (show only the matching part of a line) for pcregrep was not
1637 compatible with GNU grep in that, if there was more than one match in a
1638 line, it showed only the first of them. It now behaves in the same way as
1639 GNU grep.
1640
1641 13. If the -o and -v options were combined for pcregrep, it printed a blank
1642 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1643 does the same. The return code can be used to tell if there were any
1644 non-matching lines.
1645
1646 14. Added --file-offsets and --line-offsets to pcregrep.
1647
1648 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1649 infinitely looping recursion. The bug was that positive lookaheads were not
1650 being skipped when checking for a possible empty match (negative lookaheads
1651 and both kinds of lookbehind were skipped).
1652
1653 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1654 inclusion of <windows.h> to before rather than after the definition of
1655 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1656
1657 17. Specifying a possessive quantifier with a specific limit for a Unicode
1658 character property caused pcre_compile() to compile bad code, which led at
1659 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1660 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1661 caused the error; without that there was no problem.
1662
1663 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1664
1665 19. Added --enable-pcretest-libreadline.
1666
1667 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1668 RE::GlobalReplace(). As a result, the number of replacements returned was
1669 double what it should be. I removed one of the increments, but Craig sent a
1670 later patch that removed the other one (the right fix) and added unit tests
1671 that check the return values (which was not done before).
1672
1673 21. Several CMake things:
1674
1675 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1676 the names libpcre and libpcreposix, not just pcre and pcreposix.
1677
1678 (2) The above change means that pcretest and pcregrep are now correctly
1679 linked with the newly-built libraries, not previously installed ones.
1680
1681 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1682
1683 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1684 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1685 UTF-8 newline character). The key issue is that the pattern starts .*;
1686 this means that the match must be either at the beginning, or after a
1687 newline. The bug was in the code for advancing after a failed match and
1688 checking that the new position followed a newline. It was not taking
1689 account of UTF-8 characters correctly.
1690
1691 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1692 character classes. PCRE was not treating the sequence [:...:] as a
1693 character class unless the ... were all letters. Perl, however, seems to
1694 allow any characters between [: and :], though of course it rejects as
1695 unknown any "names" that contain non-letters, because all the known class
1696 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1697 for example, whereas PCRE did not - it did not recognize a POSIX character
1698 class. This seemed a bit dangerous, so the code has been changed to be
1699 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1700 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1701 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1702 Perl does, and where it didn't before.
1703
1704 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1705 Windows environments %n is disabled by default.
1706
1707
1708 Version 7.4 21-Sep-07
1709 ---------------------
1710
1711 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1712 means that a class such as [\s] counted as "explicit reference to CR or
1713 LF". That isn't really right - the whole point of the change was to try to
1714 help when there was an actual mention of one of the two characters. So now
1715 the change happens only if \r or \n (or a literal CR or LF) character is
1716 encountered.
1717
1718 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1719 of both had grown to the point where there were only 3 bits left.
1720 Fortunately, there was spare space in the data structure, and so I have
1721 moved the internal flags into a new 16-bit field to free up more option
1722 bits.
1723
1724 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1725 but did not set the internal JCHANGED flag - either of these is enough to
1726 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1727 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1728 start sets both bits.
1729
1730 4. Added options (at build time, compile time, exec time) to change \R from
1731 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1732
1733 5. doc/pcresyntax.html was missing from the distribution.
1734
1735 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1736 compatibility, even though it is no longer used.
1737
1738 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1739 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1740 windows.h file is present (where different names are used). [This was
1741 reversed later after testing - see 16 below.]
1742
1743 8. Changed all #include <config.h> to #include "config.h". There were also
1744 some further <pcre.h> cases that I changed to "pcre.h".
1745
1746 9. When pcregrep was used with the --colour option, it missed the line ending
1747 sequence off the lines that it output.
1748
1749 10. It was pointed out to me that arrays of string pointers cause lots of
1750 relocations when a shared library is dynamically loaded. A technique of
1751 using a single long string with a table of offsets can drastically reduce
1752 these. I have refactored PCRE in four places to do this. The result is
1753 dramatic:
1754
1755 Originally: 290
1756 After changing UCP table: 187
1757 After changing error message table: 43
1758 After changing table of "verbs" 36
1759 After changing table of Posix names 22
1760
1761 Thanks to the folks working on Gregex for glib for this insight.
1762
1763 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1764 unicode-properties was also set.
1765
1766 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1767
1768 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1769 checked only for CRLF.
1770
1771 14. Added casts to pcretest.c to avoid compiler warnings.
1772
1773 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1774
1775 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1776 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1777 entirely. This removes changes made in 7 above.
1778
1779 17. The CMake files have been updated, and there is now more information about
1780 building with CMake in the NON-UNIX-USE document.
1781
1782
1783 Version 7.3 28-Aug-07
1784 ---------------------
1785
1786 1. In the rejigging of the build system that eventually resulted in 7.1, the
1787 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1788 brackets there is not right, since it causes compilers to look for an
1789 installed pcre.h, not the version that is in the source that is being
1790 compiled (which of course may be different). I have changed it back to:
1791
1792 #include "pcre.h"
1793
1794 I have a vague recollection that the change was concerned with compiling in
1795 different directories, but in the new build system, that is taken care of
1796 by the VPATH setting the Makefile.
1797
1798 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1799 when the subject happened to end in the byte 0x85 (e.g. if the last
1800 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1801 characters but of course it shouldn't be taken as a newline when it is part
1802 of another character. The bug was that, for an unlimited repeat of . in
1803 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1804 characters when looking for a newline.
1805
1806 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1807
1808 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1809 in debug output.
1810
1811 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1812 long printing in the pcrecpp unittest when running under MinGW.
1813
1814 6. ESC_K was left out of the EBCDIC table.
1815
1816 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1817 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1818 limit also applies to "virtual nesting" when a pattern is recursive, and in
1819 this case 1000 isn't so big. I have been able to remove this limit at the
1820 expense of backing off one optimization in certain circumstances. Normally,
1821 when pcre_exec() would call its internal match() function recursively and
1822 immediately return the result unconditionally, it uses a "tail recursion"
1823 feature to save stack. However, when a subpattern that can match an empty
1824 string has an unlimited repetition quantifier, it no longer makes this
1825 optimization. That gives it a stack frame in which to save the data for
1826 checking that an empty string has been matched. Previously this was taken
1827 from the 1000-entry workspace that had been reserved. So now there is no
1828 explicit limit, but more stack is used.
1829
1830 8. Applied Daniel's patches to solve problems with the import/export magic
1831 syntax that is required for Windows, and which was going wrong for the
1832 pcreposix and pcrecpp parts of the library. These were overlooked when this
1833 problem was solved for the main library.
1834
1835 9. There were some crude static tests to avoid integer overflow when computing
1836 the size of patterns that contain repeated groups with explicit upper
1837 limits. As the maximum quantifier is 65535, the maximum group length was
1838 set at 30,000 so that the product of these two numbers did not overflow a
1839 32-bit integer. However, it turns out that people want to use groups that
1840 are longer than 30,000 bytes (though not repeat them that many times).
1841 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1842 made it possible to implement the integer overflow checks in a much more
1843 dynamic way, which I have now done. The artificial limitation on group
1844 length has been removed - we now have only the limit on the total length of
1845 the compiled pattern, which depends on the LINK_SIZE setting.
1846
1847 10. Fixed a bug in the documentation for get/copy named substring when
1848 duplicate names are permitted. If none of the named substrings are set, the
1849 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1850 empty string.
1851
1852 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1853 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1854 because the ] is interpreted as the first data character and the
1855 terminating ] is not found. PCRE has been made compatible with Perl in this
1856 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1857 cause memory overwriting.
1858
1859 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1860 string has been matched (to stop an infinite loop). It was not recognizing
1861 a conditional subpattern that could match an empty string if that
1862 subpattern was within another subpattern. For example, it looped when
1863 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1864 condition was not nested. This bug has been fixed.
1865
1866 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1867 past the start of the subject in the presence of bytes with the top bit
1868 set, for example "\x8aBCD".
1869
1870 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1871 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1872
1873 14. Optimized (?!) to (*FAIL).
1874
1875 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1876 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1877 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1878 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1879 does: it's just the validity check that is more restrictive.
1880
1881 16. Inserted checks for integer overflows during escape sequence (backslash)
1882 processing, and also fixed erroneous offset values for syntax errors during
1883 backslash processing.
1884
1885 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1886 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1887
1888 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1889 caused an overrun.
1890
1891 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1892 something other than just ASCII characters) inside a group that had an
1893 unlimited repeat caused a loop at compile time (while checking to see
1894 whether the group could match an empty string).
1895
1896 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1897 [\P{Any}] did so. (Error in the code for printing property names.)
1898
1899 21. An orphan \E inside a character class could cause a crash.
1900
1901 22. A repeated capturing bracket such as (A)? could cause a wild memory
1902 reference during compilation.
1903
1904 23. There are several functions in pcre_compile() that scan along a compiled
1905 expression for various reasons (e.g. to see if it's fixed length for look
1906 behind). There were bugs in these functions when a repeated \p or \P was
1907 present in the pattern. These operators have additional parameters compared
1908 with \d, etc, and these were not being taken into account when moving along
1909 the compiled data. Specifically:
1910
1911 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1912 length.
1913
1914 (b) An item such as \pL+ within a repeated group could cause crashes or
1915 loops.
1916
1917 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1918 "reference to non-existent subpattern" error.
1919
1920 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1921
1922 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1923 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1924
1925 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1926
1927 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1928 character were causing crashes (broken optimization).
1929
1930 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1931 \p or \P) caused a compile-time loop.
1932
1933 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1934 break. For example, the unstudied pattern [\r\n]A does not match the string
1935 "\r\nA" because change 7.0/46 below moves the current point on by two
1936 characters after failing to match at the start. However, the pattern \nA
1937 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1938 the same is true. There doesn't seem any very clean way out of this, but
1939 what I have chosen to do makes the common cases work: PCRE now takes note
1940 of whether there can be an explicit match for \r or \n anywhere in the
1941 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1942 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1943 pattern has explicit CR or LF references.
1944
1945 29. Added (*CR) etc for changing newline setting at start of pattern.
1946
1947
1948 Version 7.2 19-Jun-07
1949 ---------------------
1950
1951 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1952 which is apparently normally available under Windows.
1953
1954 2. Re-jig the pcregrep tests with different newline settings in an attempt
1955 to make them independent of the local environment's newline setting.
1956
1957 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1958
1959 4. Some of the "internals" tests were previously cut out when the link size
1960 was not 2, because the output contained actual offsets. The recent new
1961 "Z" feature of pcretest means that these can be cut out, making the tests
1962 usable with all link sizes.
1963
1964 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1965 stack recursion. This gives a massive performance boost under BSD, but just
1966 a small improvement under Linux. However, it saves one field in the frame
1967 in all cases.
1968
1969 6. Added more features from the forthcoming Perl 5.10:
1970
1971 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1972 recursion call. It refers to the nth most recently opened parentheses.
1973
1974 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1975 to be opened parentheses.
1976
1977 (c) Conditions that refer to capturing parentheses can be specified
1978 relatively, for example, (?(-2)... or (?(+3)...
1979
1980 (d) \K resets the start of the current match so that everything before
1981 is not part of it.
1982
1983 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1984
1985 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1986 reference syntax.
1987
1988 (g) (?| introduces a group in which the numbering of parentheses in each
1989 alternative starts with the same number.
1990
1991 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1992
1993 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1994 PCRE_INFO_JCHANGED.
1995
1996 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1997 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1998 for detecting groups that can match an empty string.
1999
2000 9. A pattern with a very large number of alternatives (more than several
2001 hundred) was running out of internal workspace during the pre-compile
2002 phase, where pcre_compile() figures out how much memory will be needed. A
2003 bit of new cunning has reduced the workspace needed for groups with
2004 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2005 workspace instead of running out of the 4096 that are available.
2006
2007 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2008
2009 11. Applied patch from Google to remove an optimization that didn't quite work.
2010 The report of the bug said:
2011
2012 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2013 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2014 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2015
2016 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2017 it matched the wrong number of bytes.
2018
2019
2020 Version 7.1 24-Apr-07
2021 ---------------------
2022
2023 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2024 that is more "standard", making use of automake and other Autotools. There
2025 is some re-arrangement of the files and adjustment of comments consequent
2026 on this.
2027
2028 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2029 for recursive directory scanning broke on some systems because the files
2030 are not scanned in any specific order and on different systems the order
2031 was different. A call to "sort" has been inserted into RunGrepTest for the
2032 approprate test as a short-term fix. In the longer term there may be an
2033 alternative.
2034
2035 3. I had an email from Eric Raymond about problems translating some of PCRE's
2036 man pages to HTML (despite the fact that I distribute HTML pages, some
2037 people do their own conversions for various reasons). The problems
2038 concerned the use of low-level troff macros .br and .in. I have therefore
2039 removed all such uses from the man pages (some were redundant, some could
2040 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2041 HTML has been updated to handle .nf/.fi and to complain if it encounters
2042 .br or .in.
2043
2044 4. Updated comments in configure.ac that get placed in config.h.in and also
2045 arranged for config.h to be included in the distribution, with the name
2046 config.h.generic, for the benefit of those who have to compile without
2047 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2048
2049 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2050 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2051 makevp.bat for the current PCRE, using the additional files
2052 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2053
2054 6. A Windows user reported a minor discrepancy with test 2, which turned out
2055 to be caused by a trailing space on an input line that had got lost in his
2056 copy. The trailing space was an accident, so I've just removed it.
2057
2058 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2059 that is needed.
2060
2061 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2062 as "const" (a) because they are and (b) because it helps the PHP
2063 maintainers who have recently made a script to detect big data structures
2064 in the php code that should be moved to the .rodata section. I remembered
2065 to update Builducptable as well, so it won't revert if ucptable.h is ever
2066 re-created.
2067
2068 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2069 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2070 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2071 support is not required. This saves 1.5-2K of code, which is important in
2072 some applications.
2073
2074 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2075 so as not to refer to the tables, even though these functions will never be
2076 called when UTF-8 support is disabled. Otherwise there are problems with a
2077 shared library.
2078
2079 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2080
2081 (a) It was defining its arguments as char * instead of void *.
2082
2083 (b) It was assuming that all moves were upwards in memory; this was true
2084 a long time ago when I wrote it, but is no longer the case.
2085
2086 The emulated memove() is provided for those environments that have neither
2087 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2088 is clearly not the case, as these two bugs were recently reported.
2089
2090 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2091 and Detrail to create the HTML documentation, the .txt form of the man
2092 pages, and it removes trailing spaces from listed files. It also creates
2093 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2094 case, it wraps all the #defines with #ifndefs. This script should be run
2095 before "make dist".
2096
2097 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2098 with Unicode property support.
2099
2100 (a) For a maximizing quantifier, if the two different cases of the
2101 character were of different lengths in their UTF-8 codings (there are
2102 some cases like this - I found 11), and the matching function had to
2103 back up over a mixture of the two cases, it incorrectly assumed they
2104 were both the same length.
2105
2106 (b) When PCRE was configured to use the heap rather than the stack for
2107 recursion during matching, it was not correctly preserving the data for
2108 the other case of a UTF-8 character when checking ahead for a match
2109 while processing a minimizing repeat. If the check also involved
2110 matching a wide character, but failed, corruption could cause an
2111 erroneous result when trying to check for a repeat of the original
2112 character.
2113
2114 13. Some tidying changes to the testing mechanism:
2115
2116 (a) The RunTest script now detects the internal link size and whether there
2117 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2118 values substituted by "configure". (The RunGrepTest script already did
2119 this for UTF-8.) The configure.ac script no longer substitutes the
2120 relevant variables.
2121
2122 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2123 with length and offset values. This means that the output is different
2124 for different internal link sizes. Test 2 is skipped for link sizes
2125 other than 2 because of this, bypassing the problem. Unfortunately,
2126 there was also a test in test 3 (the locale tests) that used /B and
2127 failed for link sizes other than 2. Rather than cut the whole test out,
2128 I have added a new /Z option to pcretest that replaces the length and
2129 offset values with spaces. This is now used to make test 3 independent
2130 of link size. (Test 2 will be tidied up later.)
2131
2132 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2133 segmentation fault instead of returning the appropriate error message.
2134
2135 15. In multiline mode when the newline sequence was set to "any", the pattern
2136 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2137 This doesn't seem right; it now treats the CRLF combination as the line
2138 ending, and so does not match in that case. It's only a pattern such as ^$
2139 that would hit this one: something like ^ABC$ would have failed after \r
2140 and then tried again after \r\n.
2141
2142 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2143 in an attempt to make files that differ only in their line terminators
2144 compare equal. This works on Linux.
2145
2146 17. Under certain error circumstances pcregrep might try to free random memory
2147 as it exited. This is now fixed, thanks to valgrind.
2148
2149 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2150 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2151 was because its rules for how to advance for /g after matching an empty
2152 string at the end of a line did not allow for this case. They now check for
2153 it specially.
2154
2155 20. pcretest is supposed to handle patterns and data of any length, by
2156 extending its buffers when necessary. It was getting this wrong when the
2157 buffer for a data line had to be extended.
2158
2159 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2160 CRLF as a newline sequence.
2161
2162 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2163 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2164 I have nevertheless tidied it up.
2165
2166 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2167
2168 24. Added a man page for pcre-config.
2169
2170
2171 Version 7.0 19-Dec-06
2172 ---------------------
2173
2174 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2175 moving to gcc 4.1.1.
2176
2177 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2178 sys/time.h, which is documented as needed for this function. It doesn't
2179 seem to matter on Linux, but it showed up on some releases of OS X.
2180
2181 3. It seems that there are systems where bytes whose values are greater than
2182 127 match isprint() in the "C" locale. The "C" locale should be the
2183 default when a C program starts up. In most systems, only ASCII printing
2184 characters match isprint(). This difference caused the output from pcretest
2185 to vary, making some of the tests fail. I have changed pcretest so that:
2186
2187 (a) When it is outputting text in the compiled version of a pattern, bytes
2188 other than 32-126 are always shown as hex escapes.
2189
2190 (b) When it is outputting text that is a matched part of a subject string,
2191 it does the same, unless a different locale has been set for the match
2192 (using the /L modifier). In this case, it uses isprint() to decide.
2193
2194 4. Fixed a major bug that caused incorrect computation of the amount of memory
2195 required for a compiled pattern when options that changed within the
2196 pattern affected the logic of the preliminary scan that determines the
2197 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2198 that the computed length was too small. The symptoms of this bug were
2199 either the PCRE error "internal error: code overflow" from pcre_compile(),
2200 or a glibc crash with a message such as "pcretest: free(): invalid next
2201 size (fast)". Examples of patterns that provoked this bug (shown in
2202 pcretest format) are:
2203
2204 /(?-x: )/x
2205 /(?x)(?-x: \s*#\s*)/
2206 /((?i)[\x{c0}])/8
2207 /(?i:[\x{c0}])/8
2208
2209 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2210 is now done differently.
2211
2212 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2213 wrapper classes; (b) implement a new function in the C++ scanner that is
2214 more efficient than the old way of doing things because it avoids levels of
2215 recursion in the regex matching; (c) add a paragraph to the documentation
2216 for the FullMatch() function.
2217
2218 6. The escape sequence \n was being treated as whatever was defined as
2219 "newline". Not only was this contrary to the documentation, which states
2220 that \n is character 10 (hex 0A), but it also went horribly wrong when
2221 "newline" was defined as CRLF. This has been fixed.
2222
2223 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2224 was being set to -1 for the "end of line" case (supposedly a value that no
2225 character can have). Though this value is never used (the check for end of
2226 line is "zero bytes in current character"), it caused compiler complaints.
2227 I've changed it to 0xffffffff.
2228
2229 8. In pcre_version.c, the version string was being built by a sequence of
2230 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2231 string (as it is for production releases) called a macro with an empty
2232 argument. The C standard says the result of this is undefined. The gcc
2233 compiler treats it as an empty string (which was what was wanted) but it is
2234 reported that Visual C gives an error. The source has been hacked around to
2235 avoid this problem.
2236
2237 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2238 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2239 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2240 of them did).
2241
2242 10. Originally, pcretest opened its input and output without "b"; then I was
2243 told that "b" was needed in some environments, so it was added for release
2244 5.0 to both the input and output. (It makes no difference on Unix-like
2245 systems.) Later I was told that it is wrong for the input on Windows. I've
2246 now abstracted the modes into two macros, to make it easier to fiddle with
2247 them, and removed "b" from the input mode under Windows.
2248
2249 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2250
2251 12. Added -help and --help to pcretest as an official way of being reminded
2252 of the options.
2253
2254 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2255 and pcrecpp.cc because they annoy compilers at high warning levels.
2256
2257 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2258
2259 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2260 scripts are not C programs :-) and which was not noticed because it works
2261 on Linux.
2262
2263 16. pcretest is supposed to handle any length of pattern and data line (as one
2264 line or as a continued sequence of lines) by extending its input buffer if
2265 necessary. This feature was broken for very long pattern lines, leading to
2266 a string of junk being passed to pcre_compile() if the pattern was longer
2267 than about 50K.
2268
2269 17. I have done a major re-factoring of the way pcre_compile() computes the
2270 amount of memory needed for a compiled pattern. Previously, there was code
2271 that made a preliminary scan of the pattern in order to do this. That was
2272 OK when PCRE was new, but as the facilities have expanded, it has become
2273 harder and harder to keep it in step with the real compile phase, and there
2274 have been a number of bugs (see for example, 4 above). I have now found a
2275 cunning way of running the real compile function in a "fake" mode that
2276 enables it to compute how much memory it would need, while actually only
2277 ever using a few hundred bytes of working memory and without too many
2278 tests of the mode. This should make future maintenance and development
2279 easier. A side effect of this work is that the limit of 200 on the nesting
2280 depth of parentheses has been removed (though this was never a serious
2281 limitation, I suspect). However, there is a downside: pcre_compile() now
2282 runs more slowly than before (30% or more, depending on the pattern). I
2283 hope this isn't a big issue. There is no effect on runtime performance.
2284
2285 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2286 newline (only possible for the last line of a file) and it was a
2287 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2288
2289 19. Added additional timing features to pcretest. (1) The -tm option now times
2290 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2291 separate command line item, by a number that specifies the number of
2292 repeats to use when timing. The default is 50000; this gives better
2293 precision, but takes uncomfortably long for very large patterns.
2294
2295 20. Extended pcre_study() to be more clever in cases where a branch of a
2296 subpattern has no definite first character. For example, (a*|b*)[cd] would
2297 previously give no result from pcre_study(). Now it recognizes that the
2298 first character must be a, b, c, or d.
2299
2300 21. There was an incorrect error "recursive call could loop indefinitely" if
2301 a subpattern (or the entire pattern) that was being tested for matching an
2302 empty string contained only one non-empty item after a nested subpattern.
2303 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2304 incorrectly, because the \d was being skipped in the check.
2305
2306 22. The pcretest program now has a new pattern option /B and a command line
2307 option -b, which is equivalent to adding /B to every pattern. This causes
2308 it to show the compiled bytecode, without the additional information that
2309 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2310 is the same as /B/I).
2311
2312 23. A new optimization is now able automatically to treat some sequences such
2313 as a*b as a*+b. More specifically, if something simple (such as a character
2314 or a simple class like \d) has an unlimited quantifier, and is followed by
2315 something that cannot possibly match the quantified thing, the quantifier
2316 is automatically "possessified".
2317
2318 24. A recursive reference to a subpattern whose number was greater than 39
2319 went wrong under certain circumstances in UTF-8 mode. This bug could also
2320 have affected the operation of pcre_study().
2321
2322 25. Realized that a little bit of performance could be had by replacing
2323 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2324
2325 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2326
2327 27. Possessive quantifiers such as a++ were previously implemented by turning
2328 them into atomic groups such as ($>a+). Now they have their own opcodes,
2329 which improves performance. This includes the automatically created ones
2330 from 23 above.
2331
2332 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2333 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2334 the first matched character to be a colon. This applied both to named and
2335 numbered groups.
2336
2337 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2338
2339 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2340 building PCRE on Windows easier, so I have included it in the distribution.
2341
2342 31. There is now a check in pcretest against a ridiculously large number being
2343 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2344 loop, the loop is abandoned.
2345
2346 32. Forward references to subpatterns in conditions such as (?(2)...) where
2347 subpattern 2 is defined later cause pcre_compile() to search forwards in
2348 the pattern for the relevant set of parentheses. This search went wrong
2349 when there were unescaped parentheses in a character class, parentheses
2350 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2351
2352 33. "Subroutine" calls and backreferences were previously restricted to
2353 referencing subpatterns earlier in the regex. This restriction has now
2354 been removed.
2355
2356 34. Added a number of extra features that are going to be in Perl 5.10. On the
2357 whole, these are just syntactic alternatives for features that PCRE had
2358 previously implemented using the Python syntax or my own invention. The
2359 other formats are all retained for compatibility.
2360
2361 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2362 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2363 also .NET compatible.
2364
2365 (b) A recursion or subroutine call to a named group can now be defined as
2366 (?&name) as well as (?P>name).
2367
2368 (c) A backreference to a named group can now be defined as \k<name> or
2369 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2370 5.10, are also .NET compatible.
2371
2372 (d) A conditional reference to a named group can now use the syntax
2373 (?(<name>) or (?('name') as well as (?(name).
2374
2375 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2376 groups (named and numbered) that are never evaluated inline, but can be
2377 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2378 is always false. There may be only one alternative in such a group.
2379
2380 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2381 as the simple (?(R). The condition is true only if the most recent
2382 recursion is that of the given number or name. It does not search out
2383 through the entire recursion stack.
2384
2385 (g) The escape \gN or \g{N} has been added, where N is a positive or
2386 negative number, specifying an absolute or relative reference.
2387
2388 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2389 some "unreachable code" warnings.
2390
2391 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2392 things, this adds five new scripts.
2393
2394 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2395 There were also incompatibilities regarding the handling of \Q..\E inside
2396 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2397 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2398
2399 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2400 matches an empty string, and forcibly breaks the loop. There were bugs in
2401 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2402 against aaaa the result was just "a" rather than "aaaa", for example. Two
2403 separate and independent bugs (that affected different cases) have been
2404 fixed.
2405
2406 39. Refactored the code to abolish the use of different opcodes for small
2407 capturing bracket numbers. This is a tidy that I avoided doing when I
2408 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2409 The new approach is not only tidier, it makes it possible to reduce the
2410 memory needed to fix the previous bug (38).
2411
2412 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2413 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2414 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2415 mode.
2416
2417 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2418 report.
2419
2420 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2421 copy construction and assignment in the C++ wrapper.
2422
2423 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2424 couple of bugs that could have given wrong results in the "--newline=crlf"
2425 case.
2426
2427 44. Added a number of casts and did some reorganization of signed/unsigned int
2428 variables following suggestions from Dair Grant. Also renamed the variable
2429 "this" as "item" because it is a C++ keyword.
2430
2431 45. Arranged for dftables to add
2432
2433 #include "pcre_internal.h"
2434
2435 to pcre_chartables.c because without it, gcc 4.x may remove the array
2436 definition from the final binary if PCRE is built into a static library and
2437 dead code stripping is activated.
2438
2439 46. For an unanchored pattern, if a match attempt fails at the start of a
2440 newline sequence, and the newline setting is CRLF or ANY, and the next two
2441 characters are CRLF, advance by two characters instead of one.
2442
2443
2444 Version 6.7 04-Jul-06
2445 ---------------------
2446
2447 1. In order to handle tests when input lines are enormously long, pcretest has
2448 been re-factored so that it automatically extends its buffers when
2449 necessary. The code is crude, but this _is_ just a test program. The
2450 default size has been increased from 32K to 50K.
2451
2452 2. The code in pcre_study() was using the value of the re argument before
2453 testing it for NULL. (Of course, in any sensible call of the function, it
2454 won't be NULL.)
2455
2456 3. The memmove() emulation function in pcre_internal.h, which is used on
2457 systems that lack both memmove() and bcopy() - that is, hardly ever -
2458 was missing a "static" storage class specifier.
2459
2460 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2461 containing an extended class (one that cannot be represented by a bitmap
2462 because it contains high-valued characters or Unicode property items, e.g.
2463 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2464 pattern, but PCRE should not loop if you do not (it no longer does).
2465 [Detail: two cases were found: (a) a repeated subpattern containing an
2466 extended class; (b) a recursive reference to a subpattern that followed a
2467 previous extended class. It wasn't skipping over the extended class
2468 correctly when UTF-8 mode was not set.]
2469
2470 5. A negated single-character class was not being recognized as fixed-length
2471 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2472 compile error "lookbehind assertion is not fixed length".
2473
2474 6. The RunPerlTest auxiliary script was showing an unexpected difference
2475 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2476 write a Perl script that can interpret lines of an input file either as
2477 byte characters or as UTF-8, which is what "perltest" was being required to
2478 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2479 can't do is switch easily at run time between having the "use utf8;" pragma
2480 or not. In the end, I fudged it by using the RunPerlTest script to insert
2481 "use utf8;" explicitly for the UTF-8 tests.
2482
2483 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2484 the end of the subject string, contrary to the documentation and to what
2485 Perl does. This was true of both matching functions. Now it matches only at
2486 the start of the subject and immediately after *internal* newlines.
2487
2488 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2489 a pointer to an int instead of a pointer to an unsigned long int. This
2490 caused problems on 64-bit systems.
2491
2492 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2493 instance of the 'standard' template library not being so standard".
2494
2495 10. There was no check on the number of named subpatterns nor the maximum
2496 length of a subpattern name. The product of these values is used to compute
2497 the size of the memory block for a compiled pattern. By supplying a very
2498 long subpattern name and a large number of named subpatterns, the size
2499 computation could be caused to overflow. This is now prevented by limiting
2500 the length of names to 32 characters, and the number of named subpatterns
2501 to 10,000.
2502
2503 11. Subpatterns that are repeated with specific counts have to be replicated in
2504 the compiled pattern. The size of memory for this was computed from the
2505 length of the subpattern and the repeat count. The latter is limited to
2506 65535, but there was no limit on the former, meaning that integer overflow
2507 could in principle occur. The compiled length of a repeated subpattern is
2508 now limited to 30,000 bytes in order to prevent this.
2509
2510 12. Added the optional facility to have named substrings with the same name.
2511
2512 13. Added the ability to use a named substring as a condition, using the
2513 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2514 are numbers (not recommended). Forward references are permitted.
2515
2516 14. Added forward references in named backreferences (if you see what I mean).
2517
2518 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2519 pattern could run off the end of the subject. For example, the pattern
2520 "(?s)(.{1,5})"8 did this with the subject "ab".
2521
2522 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2523 PCRE_CASELESS was set when matching characters that were quantified with ?
2524 or *.
2525
2526 17. A character class other than a single negated character that had a minimum
2527 but no maximum quantifier - for example [ab]{6,} - was not handled
2528 correctly by pce_dfa_exec(). It would match only one character.
2529
2530 18. A valid (though odd) pattern that looked like a POSIX character
2531 class but used an invalid character after [ (for example [[,abc,]]) caused
2532 pcre_compile() to give the error "Failed: internal error: code overflow" or
2533 in some cases to crash with a glibc free() error. This could even happen if
2534 the pattern terminated after [[ but there just happened to be a sequence of
2535 letters, a binary zero, and a closing ] in the memory that followed.
2536
2537 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2538 over the years. Originally (before any Unicode support), just the bottom 8
2539 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2540 output from "man perlunicode" includes this:
2541
2542 The regular expression compiler produces polymorphic opcodes. That
2543 is, the pattern adapts to the data and automatically switches to
2544 the Unicode character scheme when presented with Unicode data--or
2545 instead uses a traditional byte scheme when presented with byte
2546 data.
2547
2548 Sadly, a wide octal escape does not cause a switch, and in a string with
2549 no other multibyte characters, these octal escapes are treated as before.
2550 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2551 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2552 Unicode string.
2553
2554 I have not perpetrated such confusion in PCRE. Up till now, it took just
2555 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2556 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2557 translate to the appropriate multibyte character.
2558
2559 29. Applied some refactoring to reduce the number of warnings from Microsoft
2560 and Borland compilers. This has included removing the fudge introduced
2561 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2562 a warning about an unused variable.
2563
2564 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2565 characters since release 4.0, because Perl (from release 5.004) does not.
2566 [Or at least, is documented not to: some releases seem to be in conflict
2567 with the documentation.] However, when a pattern was studied with
2568 pcre_study() and all its branches started with \s, PCRE still included VT
2569 as a possible starting character. Of course, this did no harm; it just
2570 caused an unnecessary match attempt.
2571
2572 22. Removed a now-redundant internal flag bit that recorded the fact that case
2573 dependency changed within the pattern. This was once needed for "required
2574 byte" processing, but is no longer used. This recovers a now-scarce options
2575 bit. Also moved the least significant internal flag bit to the most-
2576 significant bit of the word, which was not previously used (hangover from
2577 the days when it was an int rather than a uint) to free up another bit for
2578 the future.
2579
2580 23. Added support for CRLF line endings as well as CR and LF. As well as the
2581 default being selectable at build time, it can now be changed at runtime
2582 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2583 specify that it is scanning data with non-default line endings.
2584
2585 24. Changed the definition of CXXLINK to make it agree with the definition of
2586 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2587
2588 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2589 recursions. This makes a big different to stack usage for some patterns.
2590
2591 26. If a subpattern containing a named recursion or subroutine reference such
2592 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2593 the space required for the compiled pattern went wrong and gave too small a
2594 value. Depending on the environment, this could lead to "Failed: internal
2595 error: code overflow at offset 49" or "glibc detected double free or
2596 corruption" errors.
2597
2598 27. Applied patches from Google (a) to support the new newline modes and (b) to
2599 advance over multibyte UTF-8 characters in GlobalReplace.
2600
2601 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2602 difference for some implementation of PCRE in some Windows version.
2603
2604 29. Added some extra testing facilities to pcretest:
2605
2606 \q<number> in a data line sets the "match limit" value
2607 \Q<number> in a data line sets the "match recursion limt" value
2608 -S <number> sets the stack size, where <number> is in megabytes
2609
2610 The -S option isn't available for Windows.
2611
2612
2613 Version 6.6 06-Feb-06
2614 ---------------------
2615
2616 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2617 in pcreposix.h. I have copied the definition from pcre.h.
2618
2619 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2620 because pcre.h is no longer a built file.
2621
2622 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2623 not normally included in the compiled code.
2624
2625
2626 Version 6.5 01-Feb-06
2627 ---------------------
2628
2629 1. When using the partial match feature with pcre_dfa_exec(), it was not
2630 anchoring the second and subsequent partial matches at the new starting
2631 point. This could lead to incorrect results. For example, with the pattern
2632 /1234/, partially matching against "123" and then "a4" gave a match.
2633
2634 2. Changes to pcregrep:
2635
2636 (a) All non-match returns from pcre_exec() were being treated as failures
2637 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2638 error message is output. Some extra information is given for the
2639 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2640 probably the only errors that are likely to be caused by users (by
2641 specifying a regex that has nested indefinite repeats, for instance).
2642 If there are more than 20 of these errors, pcregrep is abandoned.
2643
2644 (b) A binary zero was treated as data while matching, but terminated the
2645 output line if it was written out. This has been fixed: binary zeroes
2646 are now no different to any other data bytes.
2647
2648 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2649 used to set a locale for matching. The --locale=xxxx long option has
2650 been added (no short equivalent) to specify a locale explicitly on the
2651 pcregrep command, overriding the environment variables.
2652
2653 (d) When -B was used with -n, some line numbers in the output were one less
2654 than they should have been.
2655
2656 (e) Added the -o (--only-matching) option.
2657
2658 (f) If -A or -C was used with -c (count only), some lines of context were
2659 accidentally printed for the final match.
2660
2661 (g) Added the -H (--with-filename) option.
2662
2663 (h) The combination of options -rh failed to suppress file names for files
2664 that were found from directory arguments.
2665
2666 (i) Added the -D (--devices) and -d (--directories) options.
2667
2668 (j) Added the -F (--fixed-strings) option.
2669
2670 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2671
2672 (l) Added the --colo(u)r option.
2673
2674 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2675 is not present by default.
2676
2677 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2678 items such as (?R) or (?1), when the recursion could match a number of
2679 alternatives. If it matched one of the alternatives, but subsequently,
2680 outside the recursion, there was a failure, the code tried to back up into
2681 the recursion. However, because of the way PCRE is implemented, this is not
2682 possible, and the result was an incorrect result from the match.
2683
2684 In order to prevent this happening, the specification of recursion has
2685 been changed so that all such subpatterns are automatically treated as
2686 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2687
2688 4. I had overlooked the fact that, in some locales, there are characters for
2689 which isalpha() is true but neither isupper() nor islower() are true. In
2690 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2691 and ordfeminine) are like this. This affected the treatment of \w and \W
2692 when they appeared in character classes, but not when they appeared outside
2693 a character class. The bit map for "word" characters is now created
2694 separately from the results of isalnum() instead of just taking it from the
2695 upper, lower, and digit maps. (Plus the underscore character, of course.)
2696
2697 5. The above bug also affected the handling of POSIX character classes such as
2698 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2699 permanent tables. Instead, the bit maps for such a class were previously
2700 created as the appropriate unions of the upper, lower, and digit bitmaps.
2701 Now they are created by subtraction from the [[:word:]] class, which has
2702 its own bitmap.
2703
2704 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2705 It is created by subtracting the vertical space characters (\x09, \x0a,
2706 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2707 subtraction was done in the overall bitmap for a character class, meaning
2708 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2709 be recognized. This bug has been fixed.
2710
2711 7. Patches from the folks at Google:
2712
2713 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2714 real life, but is still worth protecting against".
2715
2716 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2717 regular expressions".
2718
2719 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2720 have it.
2721
2722 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2723 "configure" and the latter not, in order to fix a problem somebody had
2724 with compiling the Arg class on HP-UX.
2725
2726 (e) Improve the error-handling of the C++ wrapper a little bit.
2727
2728 (f) New tests for checking recursion limiting.
2729
2730 8. The pcre_memmove() function, which is used only if the environment does not
2731 have a standard memmove() function (and is therefore rarely compiled),
2732 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2733 returning a result (though PCRE never actually uses the result).
2734
2735 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2736 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2737 returned instead of calling malloc() with an overflowing number that would
2738 most likely cause subsequent chaos.
2739
2740 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2741
2742 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2743 with this option is matched, the nmatch and pmatch options of regexec() are
2744 ignored.
2745
2746 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2747 provided in case anyone wants to the the POSIX interface with UTF-8
2748 strings.
2749
2750 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2751 C++ linking (needed for some HP-UX environments).
2752
2753 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2754 (unused parameter) and in the pcre_printint() function (omitted "default"
2755 switch label when the default is to do nothing).
2756
2757 15. Added some code to make it possible, when PCRE is compiled as a C++
2758 library, to replace subject pointers for pcre_exec() with a smart pointer
2759 class, thus making it possible to process discontinuous strings.
2760
2761 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2762 much the same function. They were added by different people who were trying
2763 to make PCRE easy to compile on non-Unix systems. It has been suggested
2764 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2765 for compiling on Windows systems. I have therefore replaced it with
2766 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2767 defaults to "extern" for C or "extern C" for C++, which works fine on
2768 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2769 SCOPE with something explicit in config.h. In addition:
2770
2771 (a) pcreposix.h still had just "extern" instead of either of these macros;
2772 I have replaced it with PCRE_DATA_SCOPE.
2773
2774 (b) Functions such as _pcre_xclass(), which are internal to the library,
2775 but external in the C sense, all had PCRE_EXPORT in their definitions.
2776 This is apparently wrong for the Windows case, so I have removed it.
2777 (It makes no difference on Unix-like systems.)
2778
2779 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2780 of recursive calls to match(). This is different to MATCH_LIMIT because
2781 that limits the total number of calls to match(), not all of which increase
2782 the depth of recursion. Limiting the recursion depth limits the amount of
2783 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2784 when PCRE is compiled, and changed at run time. A patch from Google adds
2785 this functionality to the C++ interface.
2786
2787 18. Changes to the handling of Unicode character properties:
2788
2789 (a) Updated the table to Unicode 4.1.0.
2790
2791 (b) Recognize characters that are not in the table as "Cn" (undefined).
2792
2793 (c) I revised the way the table is implemented to a much improved format
2794 which includes recognition of ranges. It now supports the ranges that
2795 are defined in UnicodeData.txt, and it also amalgamates other
2796 characters into ranges. This has reduced the number of entries in the
2797 table from around 16,000 to around 3,000, thus reducing its size
2798 considerably. I realized I did not need to use a tree structure after
2799 all - a binary chop search is just as efficient. Having reduced the
2800 number of entries, I extended their size from 6 bytes to 8 bytes to
2801 allow for more data.
2802
2803 (d) Added support for Unicode script names via properties such as \p{Han}.
2804
2805 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2806 matching that character.
2807
2808 20. When matching a repeated Unicode property with a minimum greater than zero,
2809 (for example \pL{2,}), PCRE could look past the end of the subject if it
2810 reached it while seeking the minimum number of characters. This could
2811 happen only if some of the characters were more than one byte long, because
2812 there is a check for at least the minimum number of bytes.
2813
2814 21. Refactored the implementation of \p and \P so as to be more general, to
2815 allow for more different types of property in future. This has changed the
2816 compiled form incompatibly. Anybody with saved compiled patterns that use
2817 \p or \P will have to recompile them.
2818
2819 22. Added "Any" and "L&" to the supported property types.
2820
2821 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2822 but give a compile time error if the value is greater than 0xff.
2823
2824 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2825 accidentally not being installed or uninstalled.
2826
2827 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2828 made were to insert the current release number. This seemed silly, because
2829 it made things harder for people building PCRE on systems that don't run
2830 "configure". I have turned pcre.h into a distributed file, no longer built
2831 by "configure", with the version identification directly included. There is
2832 no longer a pcre.h.in file.
2833
2834 However, this change necessitated a change to the pcre-config script as
2835 well. It is built from pcre-config.in, and one of the substitutions was the
2836 release number. I have updated configure.ac so that ./configure now finds
2837 the release number by grepping pcre.h.
2838
2839 26. Added the ability to run the tests under valgrind.
2840
2841
2842 Version 6.4 05-Sep-05
2843 ---------------------
2844
2845 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2846 "--" to be printed when multiple files were scanned, even when none of the
2847 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2848 consider it to be a bug, and have restored the previous behaviour.
2849
2850 2. A couple of code tidies to get rid of compiler warnings.
2851
2852 3. The pcretest program used to cheat by referring to symbols in the library
2853 whose names begin with _pcre_. These are internal symbols that are not
2854 really supposed to be visible externally, and in some environments it is
2855 possible to suppress them. The cheating is now confined to including
2856 certain files from the library's source, which is a bit cleaner.
2857
2858 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2859 file's purpose clearer.
2860
2861 5. Reorganized pcre_ucp_findchar().
2862
2863
2864 Version 6.3 15-Aug-05
2865 ---------------------
2866
2867 1. The file libpcre.pc.in did not have general read permission in the tarball.
2868
2869 2. There were some problems when building without C++ support:
2870
2871 (a) If C++ support was not built, "make install" and "make test" still
2872 tried to test it.
2873
2874 (b) There were problems when the value of CXX was explicitly set. Some
2875 changes have been made to try to fix these, and ...
2876
2877 (c) --disable-cpp can now be used to explicitly disable C++ support.
2878
2879 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2880 backslash in a target when C++ was disabled. This confuses some
2881 versions of "make", apparently. Using an intermediate variable solves
2882 this. (Same for CPP_LOBJ.)
2883
2884 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2885 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2886 necessary on certain architectures.
2887
2888 4. Added a setting of -export-symbols-regex to the link command to remove
2889 those symbols that are exported in the C sense, but actually are local
2890 within the library, and not documented. Their names all begin with
2891 "_pcre_". This is not a perfect job, because (a) we have to except some
2892 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2893 available (and never for static libraries). I have made a note to try to
2894 find a way round (a) in the future.
2895
2896
2897 Version 6.2 01-Aug-05
2898 ---------------------
2899
2900 1. There was no test for integer overflow of quantifier values. A construction
2901 such as {1111111111111111} would give undefined results. What is worse, if
2902 a minimum quantifier for a parenthesized subpattern overflowed and became
2903 negative, the calculation of the memory size went wrong. This could have
2904 led to memory overwriting.
2905
2906 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2907
2908 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2909 operating environments where this matters.
2910
2911 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2912 PCRE options from within the C++ wrapper.
2913
2914 5. Named capturing subpatterns were not being correctly counted when a pattern
2915 was compiled. This caused two problems: (a) If there were more than 100
2916 such subpatterns, the calculation of the memory needed for the whole
2917 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2918 back references of the form \12, where the number was greater than 9, were
2919 not recognized as back references, even though there were sufficient
2920 previous subpatterns.
2921
2922 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2923 versions of gcc, e.g. 2.95.4.
2924
2925
2926 Version 6.1 21-Jun-05
2927 ---------------------
2928
2929 1. There was one reference to the variable "posix" in pcretest.c that was not
2930 surrounded by "#if !defined NOPOSIX".
2931
2932 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2933 the cross-check on the old pcre_info() function, for the benefit of the
2934 cut-down version of PCRE that is currently imported into Exim.
2935
2936 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2937 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2938 patterns that start (?i) but I don't think that matters. The use of (?i) is
2939 just an example; this all applies to the other options as well.
2940
2941 4. Since libtool seems to echo the compile commands it is issuing, the output
2942 from "make" can be reduced a bit by putting "@" in front of each libtool
2943 compile command.
2944
2945 5. Patch from the folks at Google for configure.in to be a bit more thorough
2946 in checking for a suitable C++ installation before trying to compile the
2947 C++ stuff. This should fix a reported problem when a compiler was present,
2948 but no suitable headers.
2949
2950 6. The man pages all had just "PCRE" as their title. I have changed them to
2951 be the relevant file name. I have also arranged that these names are
2952 retained in the file doc/pcre.txt, which is a concatenation in text format
2953 of all the man pages except the little individual ones for each function.
2954
2955 7. The NON-UNIX-USE file had not been updated for the different set of source
2956 files that come with release 6. I also added a few comments about the C++
2957 wrapper.
2958
2959
2960 Version 6.0 07-Jun-05
2961 ---------------------
2962
2963 1. Some minor internal re-organization to help with my DFA experiments.
2964
2965 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2966 didn't matter for the library itself when fully configured, but did matter
2967 when compiling without UCP support, or within Exim, where the ucp files are
2968 not imported.
2969
2970 3. Refactoring of the library code to split up the various functions into
2971 different source modules. The addition of the new DFA matching code (see
2972 below) to a single monolithic source would have made it really too
2973 unwieldy, quite apart from causing all the code to be include in a
2974 statically linked application, when only some functions are used. This is
2975 relevant even without the DFA addition now that patterns can be compiled in
2976 one application and matched in another.
2977
2978 The downside of splitting up is that there have to be some external
2979 functions and data tables that are used internally in different modules of
2980 the library but which are not part of the API. These have all had their
2981 names changed to start with "_pcre_" so that they are unlikely to clash
2982 with other external names.
2983
2984 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2985 a different (DFA) algorithm. Although it is slower than the original
2986 function, it does have some advantages for certain types of matching
2987 problem.
2988
2989 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2990 including restarting after a partial match.
2991
2992 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2993 defined when compiling for Windows was sent to me. I have put it into the
2994 code, though I have no means of testing or verifying it.
2995
2996 7. Added the pcre_refcount() auxiliary function.
2997
2998 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2999 match before or at the first newline in the subject string. In pcretest,
3000 the /f option on a pattern can be used to set this.
3001
3002 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3003 would behave wrongly. This has been present in PCRE since release 4.0.
3004
3005 10. A number of changes to the pcregrep command:
3006
3007 (a) Refactored how -x works; insert ^(...)$ instead of setting
3008 PCRE_ANCHORED and checking the length, in preparation for adding
3009 something similar for -w.
3010
3011 (b) Added the -w (match as a word) option.
3012
3013 (c) Refactored the way lines are read and buffered so as to have more
3014 than one at a time available.
3015
3016 (d) Implemented a pcregrep test script.
3017
3018 (e) Added the -M (multiline match) option. This allows patterns to match
3019 over several lines of the subject. The buffering ensures that at least
3020 8K, or the rest of the document (whichever is the shorter) is available
3021 for matching (and similarly the previous 8K for lookbehind assertions).
3022
3023 (f) Changed the --help output so that it now says
3024
3025 -w, --word-regex(p)
3026
3027 instead of two lines, one with "regex" and the other with "regexp"
3028 because that confused at least one person since the short forms are the
3029 same. (This required a bit of code, as the output is generated
3030 automatically from a table. It wasn't just a text change.)
3031
3032 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3033 option but starts with a hyphen. Could be a pattern or a path name
3034 starting with a hyphen, for instance.
3035
3036 (h) "-" can be given as a file name to represent stdin.
3037
3038 (i) When file names are being printed, "(standard input)" is used for
3039 the standard input, for compatibility with GNU grep. Previously
3040 "<stdin>" was used.
3041
3042 (j) The option --label=xxx can be used to supply a name to be used for
3043 stdin when file names are being printed. There is no short form.
3044
3045 (k) Re-factored the options decoding logic because we are going to add
3046 two more options that take data. Such options can now be given in four
3047 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3048
3049 (l) Added the -A, -B, and -C options for requesting that lines of context
3050 around matches be printed.
3051
3052 (m) Added the -L option to print the names of files that do not contain
3053 any matching lines, that is, the complement of -l.
3054
3055 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3056 continue to scan other files.
3057
3058 (o) The -s option was incorrectly implemented. For compatibility with other
3059 greps, it now suppresses the error message for a non-existent or non-
3060 accessible file (but not the return code). There is a new option called
3061 -q that suppresses the output of matching lines, which was what -s was
3062 previously doing.
3063
3064 (p) Added --include and --exclude options to specify files for inclusion
3065 and exclusion when recursing.
3066
3067 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3068 Hopefully, it now does.
3069
3070 12. Missing cast in pcre_study().
3071
3072 13. Added an "uninstall" target to the makefile.
3073
3074 14. Replaced "extern" in the function prototypes in Makefile.in with
3075 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3076 world, but is set differently for Windows.
3077
3078 15. Added a second compiling function called pcre_compile2(). The only
3079 difference is that it has an extra argument, which is a pointer to an
3080 integer error code. When there is a compile-time failure, this is set
3081 non-zero, in addition to the error test pointer being set to point to an
3082 error message. The new argument may be NULL if no error number is required
3083 (but then you may as well call pcre_compile(), which is now just a
3084 wrapper). This facility is provided because some applications need a
3085 numeric error indication, but it has also enabled me to tidy up the way
3086 compile-time errors are handled in the POSIX wrapper.
3087
3088 16. Added VPATH=.libs to the makefile; this should help when building with one
3089 prefix path and installing with another. (Or so I'm told by someone who
3090 knows more about this stuff than I do.)
3091
3092 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3093 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3094 match everything, including newlines. This is not POSIX-compatible, but
3095 somebody wanted the feature. From pcretest it can be activated by using
3096 both the P and the s flags.
3097
3098 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3099
3100 19. libpcre.pc was being incorrectly installed as executable.
3101
3102 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3103 it now also looks for '\r' so that it will work unmodified on Windows.
3104
3105 21. Added Google's contributed C++ wrapper to the distribution.
3106
3107 22. Added some untidy missing memory free() calls in pcretest, to keep
3108 Electric Fence happy when testing.
3109
3110
3111
3112 Version 5.0 13-Sep-04
3113 ---------------------
3114
3115 1. Internal change: literal characters are no longer packed up into items
3116 containing multiple characters in a single byte-string. Each character
3117 is now matched using a separate opcode. However, there may be more than one
3118 byte in the character in UTF-8 mode.
3119
3120 2. The pcre_callout_block structure has two new fields: pattern_position and
3121 next_item_length. These contain the offset in the pattern to the next match
3122 item, and its length, respectively.
3123
3124 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3125 insertion of callouts before each pattern item. Added the /C option to
3126 pcretest to make use of this.
3127
3128 4. On the advice of a Windows user, the lines
3129
3130 #if defined(_WIN32) || defined(WIN32)
3131 _setmode( _fileno( stdout ), 0x8000 );
3132 #endif /* defined(_WIN32) || defined(WIN32) */
3133
3134 have been added to the source of pcretest. This apparently does useful
3135 magic in relation to line terminators.
3136
3137 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3138 for the benefit of those environments where the "b" makes a difference.
3139
3140 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3141 to know about it. I have put a hack into configure.in that adds in code
3142 to set GCC=yes if CC=icc. This seems to end up at a point in the
3143 generated configure script that is early enough to affect the setting of
3144 compiler options, which is what is needed, but I have no means of testing
3145 whether it really works. (The user who reported this had patched the
3146 generated configure script, which of course I cannot do.)
3147
3148 LATER: After change 22 below (new libtool files), the configure script
3149 seems to know about icc (and also ecc). Therefore, I have commented out
3150 this hack in configure.in.
3151
3152 7. Added support for pkg-config (2 patches were sent in).
3153
3154 8. Negated POSIX character classes that used a combination of internal tables
3155 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3156 [[:^ascii]]. Typically, they would match almost any characters. The other
3157 POSIX classes were not broken in this way.
3158
3159 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3160 to find the match, as PCRE was deluded into thinking that the match had to
3161 start at the start point or following a newline. The same bug applied to
3162 patterns with negative forward assertions or any backward assertions
3163 preceding ".*" at the start, unless the pattern required a fixed first
3164 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3165
3166 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3167 starting at the last subject character, bytes beyond the end of the subject
3168 string were read.
3169
3170 11. Renamed the variable "class" as "classbits" to make life easier for C++
3171 users. (Previously there was a macro definition, but it apparently wasn't
3172 enough.)
3173
3174 12. Added the new field "tables" to the extra data so that tables can be passed
3175 in at exec time, or the internal tables can be re-selected. This allows
3176 a compiled regex to be saved and re-used at a later time by a different
3177 program that might have everything at different addresses.
3178
3179 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3180 -R library as well as a -L library.
3181
3182 14. The debugging options of pcretest (-d on the command line or D on a
3183 pattern) showed incorrect output for anything following an extended class
3184 that contained multibyte characters and which was followed by a quantifier.
3185
3186 15. Added optional support for general category Unicode character properties
3187 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3188 support. It adds about 90K to the size of the library. The meanings of the
3189 inbuilt class escapes such as \d and \s have NOT been changed.
3190
3191 16. Updated pcredemo.c to include calls to free() to release the memory for the
3192 compiled pattern.
3193
3194 17. The generated file chartables.c was being created in the source directory
3195 instead of in the building directory. This caused the build to fail if the
3196 source directory was different from the building directory, and was
3197 read-only.
3198
3199 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3200 file. No doubt somebody will tell me if they don't make sense... Also added
3201 Dan Mooney's comments about building on OpenVMS.
3202
3203 19. Added support for partial matching via the PCRE_PARTIAL option for
3204 pcre_exec() and the \P data escape in pcretest.
3205
3206 20. Extended pcretest with 3 new pattern features:
3207
3208 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3209 write the compiled pattern to the file whose name is "rest-of-line".
3210 This is a straight binary dump of the data, with the saved pointer to
3211 the character tables forced to be NULL. The study data, if any, is
3212 written too. After writing, pcretest reads a new pattern.
3213
3214 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3215 compiled pattern from the given file. There must not be any
3216 occurrences of "<" in the file name (pretty unlikely); if there are,
3217 pcretest will instead treat the initial "<" as a pattern delimiter.
3218 After reading in the pattern, pcretest goes on to read data lines as
3219 usual.
3220
3221 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3222 and 16-bit fields in a compiled pattern, to simulate a pattern that
3223 was compiled on a host of opposite endianness.
3224
3225 21. The pcre-exec() function can now cope with patterns that were compiled on
3226 hosts of opposite endianness, with this restriction:
3227
3228 As for any compiled expression that is saved and used later, the tables
3229 pointer field cannot be preserved; the extra_data field in the arguments
3230 to pcre_exec() should be used to pass in a tables address if a value
3231 other than the default internal tables were used at compile time.
3232
3233 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3234 now diagnosed as an error. Previously, most of the time, a negative number
3235 would have been treated as zero, but if in addition "ovector" was passed as
3236 NULL, a crash could occur.
3237
3238 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3239 new versions from the libtool 1.5 distribution (the last one is a copy of
3240 a file called libtool.m4). This seems to have fixed the need to patch
3241 "configure" to support Darwin 1.3 (which I used to do). However, I still
3242 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3243 workstation).
3244
3245 24. Changed the PCRE licence to be the more standard "BSD" licence.
3246
3247
3248 Version 4.5 01-Dec-03
3249 ---------------------
3250
3251 1. There has been some re-arrangement of the code for the match() function so
3252 that it can be compiled in a version that does not call itself recursively.
3253 Instead, it keeps those local variables that need separate instances for
3254 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3255 needs to "recurse". Keeping track of where control must go is done by means
3256 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3257 hide most of the details from the main code, and operates only if
3258 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3259 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3260 operating.
3261
3262 To make it easier for callers to provide specially tailored get/free
3263 functions for this usage, two new functions, pcre_stack_malloc, and
3264 pcre_stack_free, are used. They are always called in strict stacking order,
3265 and the size of block requested is always the same.
3266
3267 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3268 PCRE has been compiled to use the stack or the heap for recursion. The
3269 -C option of pcretest uses this to show which version is compiled.
3270
3271 A new data escape \S, is added to pcretest; it causes the amounts of store
3272 obtained and freed by both kinds of malloc/free at match time to be added
3273 to the output.
3274
3275 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3276 what's available on my current Linux desktop machine.
3277
3278 3. When matching a UTF-8 string, the test for a valid string at the start has
3279 been extended. If start_offset is not zero, PCRE now checks that it points
3280 to a byte that is the start of a UTF-8 character. If not, it returns
3281 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3282 this is necessary because there may be backward assertions in the pattern.
3283 When matching the same subject several times, it may save resources to use
3284 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3285
3286 4. The code for checking the validity of UTF-8 strings has been tightened so
3287 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3288 containing "overlong sequences".
3289
3290 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3291 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3292 should have been "&", but it just so happened that all the cases this let
3293 through by mistake were picked up later in the function.
3294
3295 6. I had used a variable called "isblank" - this is a C99 function, causing
3296 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3297
3298 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3299 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3300 so the version is shown; (c) stop "make test" echoing "./RunTest".
3301
3302 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3303
3304 9. The prototype for memmove() for systems that don't have it was using
3305 size_t, but the inclusion of the header that defines size_t was later. I've
3306 moved the #includes for the C headers earlier to avoid this.
3307
3308 10. Added some adjustments to the code to make it easier to compiler on certain
3309 special systems:
3310
3311 (a) Some "const" qualifiers were missing.
3312 (b) Added the macro EXPORT before all exported functions; by default this
3313 is defined to be empty.
3314 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3315 that it reads its output file name as an argument instead of writing
3316 to the standard output and assuming this can be redirected.
3317
3318 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3319 class containing characters with values greater than 255, PCRE compilation
3320 went into a loop.
3321
3322 12. A recursive reference to a subpattern that was within another subpattern
3323 that had a minimum quantifier of zero caused PCRE to crash. For example,
3324 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3325 recursion. If the recursively-called subpattern itself had a zero repeat,
3326 that was OK.
3327
3328 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3329 buffer into which it was copied (for escape processing) was still set at
3330 1024, so long lines caused crashes.
3331
3332 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3333 "internal error: code overflow...". This applied to any character class
3334 that was followed by a possessive quantifier.
3335
3336 15. Modified the Makefile to add libpcre.la as a prerequisite for
3337 libpcreposix.la because I was told this is needed for a parallel build to
3338 work.
3339
3340 16. If a pattern that contained .* following optional items at the start was
3341 studied, the wrong optimizing data was generated, leading to matching
3342 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3343 matching string must start with a or b or c. The correct conclusion for
3344 this pattern is that a match can start with any character.
3345
3346
3347 Version 4.4 13-Aug-03
3348 ---------------------
3349
3350 1. In UTF-8 mode, a character class containing characters with values between
3351 127 and 255 was not handled correctly if the compiled pattern was studied.
3352 In fixing this, I have also improved the studying algorithm for such
3353 classes (slightly).
3354
3355 2. Three internal functions had redundant arguments passed to them. Removal
3356 might give a very teeny performance improvement.
3357
3358 3. Documentation bug: the value of the capture_top field in a callout is *one
3359 more than* the number of the hightest numbered captured substring.
3360
3361 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3362 in incorrectly linking with a previously installed version. They now link
3363 explicitly with libpcre.la.
3364
3365 5. configure.in no longer needs to recognize Cygwin specially.
3366
3367 6. A problem in pcre.in for Windows platforms is fixed.
3368
3369 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3370 pcretest, it used to include the size of the study block as part of its
3371 output. Unfortunately, the structure contains a field that has a different
3372 size on different hardware architectures. This meant that the tests that
3373 showed this size failed. As the block is currently always of a fixed size,
3374 this information isn't actually particularly useful in pcretest output, so
3375 I have just removed it.
3376
3377 8. Three pre-processor statements accidentally did not start in column 1.
3378 Sadly, there are *still* compilers around that complain, even though
3379 standard C has not required this for well over a decade. Sigh.
3380
3381 9. In pcretest, the code for checking callouts passed small integers in the
3382 callout_data field, which is a void * field. However, some picky compilers
3383 complained about the casts involved for this on 64-bit systems. Now
3384 pcretest passes the address of the small integer instead, which should get
3385 rid of the warnings.
3386
3387 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3388 both compile and run time, and gives an error if an invalid UTF-8 sequence
3389 is found. There is a option for disabling this check in cases where the
3390 string is known to be correct and/or the maximum performance is wanted.
3391
3392 11. In response to a bug report, I changed one line in Makefile.in from
3393
3394 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3395 to
3396 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3397
3398 to look similar to other lines, but I have no way of telling whether this
3399 is the right thing to do, as I do not use Windows. No doubt I'll get told
3400 if it's wrong...
3401
3402
3403 Version 4.3 21-May-03
3404 ---------------------
3405
3406 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3407 Makefile.
3408
3409 2. Some refactoring to improve the quality of the code:
3410
3411 (i) The utf8_table... variables are now declared "const".
3412
3413 (ii) The code for \cx, which used the "case flipping" table to upper case
3414 lower case letters, now just substracts 32. This is ASCII-specific,
3415 but the whole concept of \cx is ASCII-specific, so it seems
3416 reasonable.
3417
3418 (iii) PCRE was using its character types table to recognize decimal and
3419 hexadecimal digits in the pattern. This is silly, because it handles
3420 only 0-9, a-f, and A-F, but the character types table is locale-
3421 specific, which means strange things might happen. A private
3422 table is now used for this - though it costs 256 bytes, a table is
3423 much faster than multiple explicit tests. Of course, the standard
3424 character types table is still used for matching digits in subject
3425 strings against \d.
3426
3427 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3428 ending in _t are). So I've renamed it as ESC_tee.
3429
3430 3. The first argument for regexec() in the POSIX wrapper should have been
3431 defined as "const".
3432
3433 4. Changed pcretest to use malloc() for its buffers so that they can be
3434 Electric Fenced for debugging.
3435
3436 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3437 to read one or more bytes before the start of the subject string. Often this
3438 had no effect on PCRE's behaviour, but in some circumstances it could
3439 provoke a segmentation fault.
3440
3441 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3442 to try to read one or more bytes before the start of the subject string.
3443
3444 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3445 UTF-8 support could misbehave in various ways if the subject string
3446 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3447 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3448 back over UTF-8 characters.)
3449
3450
3451 Version 4.2 14-Apr-03
3452 ---------------------
3453
3454 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3455
3456 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3457 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3458 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3459 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3460 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3461 and BUILD_EXEEXT
3462 Note: automatic setting of the BUILD variables is not yet working
3463 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3464 compile-time but not at link-time
3465 [LINK]: use for linking executables only
3466 make different versions for Windows and non-Windows
3467 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3468 libraries
3469 [LINK_FOR_BUILD]: new variable
3470 [OBJEXT]: use throughout
3471 [EXEEXT]: use throughout
3472 <winshared>: new target
3473 <wininstall>: new target
3474 <dftables.o>: use native compiler
3475 <dftables>: use native linker
3476 <install>: handle Windows platform correctly
3477 <clean>: ditto
3478 <check>: ditto
3479 copy DLL to top builddir before testing
3480
3481 As part of these changes, -no-undefined was removed again. This was reported
3482 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3483 in any case.
3484
3485 3. Some tidies to get rid of compiler warnings:
3486
3487 . In the match_data structure, match_limit was an unsigned long int, whereas
3488 match_call_count was an int. I've made them both unsigned long ints.
3489
3490 . In pcretest the fact that a const uschar * doesn't automatically cast to
3491 a void * provoked a warning.
3492
3493 . Turning on some more compiler warnings threw up some "shadow" variables
3494 and a few more missing casts.
3495
3496 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3497 option, a class that contained a single character with a value between 128
3498 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3499
3500 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3501 option, a class that contained several characters, but with at least one
3502 whose value was between 128 and 255 caused PCRE to crash.
3503
3504
3505 Version 4.1 12-Mar-03
3506 ---------------------
3507
3508 1. Compiling with gcc -pedantic found a couple of places where casts were
3509 needed, and a string in dftables.c that was longer than standard compilers are
3510 required to support.
3511
3512 2. Compiling with Sun's compiler found a few more places where the code could
3513 be tidied up in order to avoid warnings.
3514
3515 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3516 first of these names is deprecated in the latest Autoconf in favour of the name
3517 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3518 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3519 analogy I have changed it to CFLAGS_FOR_BUILD.
3520
3521 4. Added -no-undefined to the linking command in the Makefile, because this is
3522 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3523 linking step for the pcreposix library.
3524
3525 5. PCRE was failing to diagnose the case of two named groups with the same
3526 name.
3527
3528 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3529 literal character that is needed in the subject for a match, and scans along to
3530 ensure that it is present before embarking on the full matching process. This
3531 saves time in cases of nested unlimited repeats that are never going to match.
3532 Problem: the scan can take a lot of time if the subject is very long (e.g.
3533 megabytes), thus penalizing straightforward matches. It is now done only if the
3534 amount of subject to be scanned is less than 1000 bytes.
3535
3536 7. A lesser problem with the same optimization is that it was recording the
3537 first character of an anchored pattern as "needed", thus provoking a search
3538 right along the subject, even when the first match of the pattern was going to
3539 fail. The "needed" character is now not set for anchored patterns, unless it
3540 follows something in the pattern that is of non-fixed length. Thus, it still
3541 fulfils its original purpose of finding quick non-matches in cases of nested
3542 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3543
3544
3545 Version 4.0 17-Feb-03
3546 ---------------------
3547
3548 1. If a comment in an extended regex that started immediately after a meta-item
3549 extended to the end of string, PCRE compiled incorrect data. This could lead to
3550 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3551
3552 2. Moved to autoconf 2.53 and libtool 1.4.2.
3553
3554 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3555 the special perltest8 script is no longer needed - all the tests can be run
3556 from a single perltest script.
3557
3558 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3559 by \s. It has now been removed in PCRE. This means it isn't recognized as
3560 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3561 class [:space:] *does* include VT, thereby creating a mess.
3562
3563 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3564 space and tab.
3565
3566 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3567 its new features into the main test script, reducing the number of scripts.
3568
3569 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3570 were backward compatible, and made the (?i) apply to the whole pattern, as if
3571 /i were given. Now it behaves more logically, and applies the option setting
3572 only to what follows. PCRE has been changed to follow suit. However, if it
3573 finds options settings right at the start of the pattern, it extracts them into
3574 the global options, as before. Thus, they show up in the info data.
3575
3576 8. Added support for the \Q...\E escape sequence. Characters in between are
3577 treated as literals. This is slightly different from Perl in that $ and @ are
3578 also handled as literals inside the quotes. In Perl, they will cause variable
3579 interpolation. Note the following examples:
3580
3581 Pattern PCRE matches Perl matches
3582
3583 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3584 \Qabc\$xyz\E abc\$xyz abc\$xyz
3585 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3586
3587 For compatibility with Perl, \Q...\E sequences are recognized inside character
3588 classes as well as outside them.
3589
3590 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3591 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3592 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3593 signed/unsigned warnings.
3594
3595 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3596 option for pcretest, so I've replaced it by a simple function that does just
3597 that job.
3598
3599 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3600 "pcregrep -".
3601
3602 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3603 Java package. This provides some syntactic sugar for simple cases of what my
3604 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3605 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3606 item, you can use this simplified notation. Note that only makes sense with
3607 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3608 greediness, whatever the setting of the PCRE_UNGREEDY option.
3609
3610 13. A change of greediness default within a pattern was not taking effect at
3611 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3612 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3613 was abstracted outside.
3614
3615 14. PCRE now supports the \G assertion. It is true when the current matching
3616 position is at the start point of the match. This differs from \A when the
3617 starting offset is non-zero. Used with the /g option of pcretest (or similar
3618 code), it works in the same way as it does for Perl's /g option. If all
3619 alternatives of a regex begin with \G, the expression is anchored to the start
3620 match position, and the "anchored" flag is set in the compiled expression.
3621
3622 15. Some bugs concerning the handling of certain option changes within patterns
3623 have been fixed. These applied to options other than (?ims). For example,
3624 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3625 been the other way round. Some of this was related to change 7 above.
3626
3627 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3628 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3629 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3630 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3631
3632 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3633 mode. Unlike ".", it always matches newline, whatever the setting of
3634 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3635 assertions. Perl allows it, but it doesn't (in general) work because it can't
3636 calculate the length of the lookbehind. At least, that's the case for Perl
3637 5.8.0 - I've been told they are going to document that it doesn't work in
3638 future.
3639
3640 18. Added an error diagnosis for escapes that PCRE does not support: these are
3641 \L, \l, \N, \P, \p, \U, \u, and \X.
3642
3643 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3644 reading past the end of the pattern in cases such as /[abcd/.
3645
3646 20. PCRE was getting more memory than necessary for patterns with classes that
3647 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3648
3649 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3650 compiling PCRE for use with Virtual Pascal.
3651
3652 22. Small fix to the Makefile to make it work properly if the build is done
3653 outside the source tree.
3654
3655 23. Added a new extension: a condition to go with recursion. If a conditional
3656 subpattern starts with (?(R) the "true" branch is used if recursion has
3657 happened, whereas the "false" branch is used only at the top level.
3658
3659 24. When there was a very long string of literal characters (over 255 bytes
3660 without UTF support, over 250 bytes with UTF support), the computation of how
3661 much memory was required could be incorrect, leading to segfaults or other
3662 strange effects.
3663
3664 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3665 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3666 there was a subsequent back reference to those brackets. This meant that, for
3667 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3668 possible to check for precisely this case. All we can do is abandon the
3669 optimization if .* occurs inside capturing brackets when there are any back
3670 references whatsoever. (See below for a better fix that came later.)
3671
3672 26. The handling of the optimization for finding the first character of a
3673 non-anchored pattern, and for finding a character that is required later in the
3674 match were failing in some cases. This didn't break the matching; it just
3675 failed to optimize when it could. The way this is done has been re-implemented.
3676
3677 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3678
3679 28. Added a new feature that provides some of the functionality that Perl
3680 provides with (?{...}). The facility is termed a "callout". The way it is done
3681 in PCRE is for the caller to provide an optional function, by setting
3682 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3683 global variable. By default it is unset, which disables all calling out. To get
3684 the function called, the regex must include (?C) at appropriate points. This
3685 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3686 This provides a means of identifying different callout points. When PCRE
3687 reaches such a point in the regex, if pcre_callout has been set, the external
3688 function is called. It is provided with data in a structure called
3689 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3690 matching continues; if it returns a non-zero value, the match at the current
3691 point fails. However, backtracking will occur if possible. [This was changed
3692 later and other features added - see item 49 below.]
3693
3694 29. pcretest is upgraded to test the callout functionality. It provides a
3695 callout function that displays information. By default, it shows the start of
3696 the match and the current position in the text. There are some new data escapes
3697 to vary what happens:
3698
3699 \C+ in addition, show current contents of captured substrings
3700 \C- do not supply a callout function
3701 \C!n return 1 when callout number n is reached
3702 \C!n!m return 1 when callout number n is reached for the mth time
3703
3704 30. If pcregrep was called with the -l option and just a single file name, it
3705 output "<stdin>" if a match was found, instead of the file name.
3706
3707 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3708 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3709 pcre_exec(). This saves a malloc/free per call. The default value of
3710 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3711 when configuring.
3712
3713 32. The default maximum size of a compiled pattern is 64K. There have been a
3714 few cases of people hitting this limit. The code now uses macros to handle the
3715 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3716 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3717 configuring. Tests 2 and 5 work only with 2-byte links because they output
3718 debugging information about compiled patterns.
3719
3720 33. Internal code re-arrangements:
3721
3722 (a) Moved the debugging function for printing out a compiled regex into
3723 its own source file (printint.c) and used #include to pull it into
3724 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3725 separate copies.
3726
3727 (b) Defined the list of op-code names for debugging as a macro in
3728 internal.h so that it is next to the definition of the opcodes.
3729
3730 (c) Defined a table of op-code lengths for simpler skipping along compiled
3731 code. This is again a macro in internal.h so that it is next to the
3732 definition of the opcodes.
3733
3734 34. Added support for recursive calls to individual subpatterns, along the
3735 lines of Robin Houston's patch (but implemented somewhat differently).
3736
3737 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3738 allow it to read and process whole directories in Win32. This code was
3739 contributed by Lionel Fourquaux; it has not been tested by me.
3740
3741 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3742 used to name a group. Names consist of alphanumerics and underscores, and must
3743 be unique. Back references use the syntax (?P=name) and recursive calls use
3744 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3745 numbers. The function pcre_fullinfo() can be used after compilation to extract
3746 a name/number map. There are three relevant calls:
3747
3748 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3749 PCRE_INFO_NAMECOUNT yields the number of entries
3750 PCRE_INFO_NAMETABLE yields a pointer to the map.
3751
3752 The map is a vector of fixed-size entries. The size of each entry depends on
3753 the length of the longest name used. The first two bytes of each entry are the
3754 group number, most significant byte first. There follows the corresponding
3755 name, zero terminated. The names are in alphabetical order.
3756
3757 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3758 case instead of 255. Making it the same both with and without UTF-8 support
3759 means that the same test output works with both.
3760
3761 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3762 calling malloc() with a zero argument.
3763
3764 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3765 optimization. I've improved things by keeping a bitmap of backreferences with
3766 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3767 fact referenced, the optimization can be applied. It is unlikely that a
3768 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3769 the match to follow \n) will appear inside brackets with a number greater than
3770 31, but if it does, any back reference > 31 suppresses the optimization.
3771
3772 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3773 of disabling numbered capturing parentheses. Any opening parenthesis that is
3774 not followed by ? behaves as if it were followed by ?: but named parentheses
3775 can still be used for capturing (and they will acquire numbers in the usual
3776 way).
3777
3778 41. Redesigned the return codes from the match() function into yes/no/error so
3779 that errors can be passed back from deep inside the nested calls. A malloc
3780 failure while inside a recursive subpattern call now causes the
3781 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3782
3783 42. It is now possible to set a limit on the number of times the match()
3784 function is called in a call to pcre_exec(). This facility makes it possible to
3785 limit the amount of recursion and backtracking, though not in a directly
3786 obvious way, because the match() function is used in a number of different
3787 circumstances. The count starts from zero for each position in the subject
3788 string (for non-anchored patterns). The default limit is, for compatibility, a
3789 large number, namely 10 000 000. You can change this in two ways:
3790
3791 (a) When configuring PCRE before making, you can use --with-match-limit=n
3792 to set a default value for the compiled library.
3793
3794 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3795 a different value is set. See 45 below.
3796
3797 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3798
3799 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3800 of things that can be changed at compile time. The first argument specifies
3801 what is wanted and the second points to where the information is to be placed.
3802 The current list of available information is:
3803
3804 PCRE_CONFIG_UTF8
3805
3806 The output is an integer that is set to one if UTF-8 support is available;
3807 otherwise it is set to zero.
3808
3809 PCRE_CONFIG_NEWLINE
3810
3811 The output is an integer that it set to the value of the code that is used for
3812 newline. It is either LF (10) or CR (13).
3813
3814 PCRE_CONFIG_LINK_SIZE
3815
3816 The output is an integer that contains the number of bytes used for internal
3817 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3818
3819 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3820
3821 The output is an integer that contains the threshold above which the POSIX
3822 interface uses malloc() for output vectors. See item 31 above.
3823
3824 PCRE_CONFIG_MATCH_LIMIT
3825
3826 The output is an unsigned integer that contains the default limit of the number
3827 of match() calls in a pcre_exec() execution. See 42 above.
3828
3829 44. pcretest has been upgraded by the addition of the -C option. This causes it
3830 to extract all the available output from the new pcre_config() function, and to
3831 output it. The program then exits immediately.
3832
3833 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3834 order to support additional features. One way would have been to define
3835 pcre_exec2() (for example) with extra arguments, but this would not have been
3836 extensible, and would also have required all calls to the original function to
3837 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3838 is used for passing in "extra" data from pcre_study().
3839
3840 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3841 contains the following fields:
3842
3843 flags a bitmap indicating which of the following fields are set
3844 study_data opaque data from pcre_study()
3845 match_limit a way of specifying a limit on match() calls for a specific
3846 call to pcre_exec()
3847 callout_data data for callouts (see 49 below)
3848
3849 The flag bits are also defined in pcre.h, and are
3850
3851 PCRE_EXTRA_STUDY_DATA
3852 PCRE_EXTRA_MATCH_LIMIT
3853 PCRE_EXTRA_CALLOUT_DATA
3854
3855 The pcre_study() function now returns one of these new pcre_extra blocks, with
3856 the actual study data pointed to by the study_data field, and the
3857 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3858 before. That is, this change is entirely upwards-compatible and requires no
3859 change to existing code.
3860
3861 If you want to pass in additional data to pcre_exec(), you can either place it
3862 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3863 block.
3864
3865 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3866 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3867 times with different match limits, until it finds the minimum value needed for
3868 pcre_exec() to complete. The value is then output. This can be instructive; for
3869 most simple matches the number is quite small, but for pathological cases it
3870 gets very large very quickly.
3871
3872 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3873 returns the size of the data block pointed to by the study_data field in a
3874 pcre_extra block, that is, the value that was passed as the argument to
3875 pcre_malloc() when PCRE was getting memory in which to place the information
3876 created by pcre_study(). The fourth argument should point to a size_t variable.
3877 pcretest has been extended so that this information is shown after a successful
3878 pcre_study() call when information about the compiled regex is being displayed.
3879
3880 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3881 because what follows is always an absolute path. (Later: it turns out that this
3882 is more than cosmetic for MinGW, because it doesn't like empty path
3883 components.)
3884
3885 49. Some changes have been made to the callout feature (see 28 above):
3886
3887 (i) A callout function now has three choices for what it returns:
3888
3889 0 => success, carry on matching
3890 > 0 => failure at this point, but backtrack if possible
3891 < 0 => serious error, return this value from pcre_exec()
3892
3893 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3894 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3895 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3896 use by callout functions. It will never be used by PCRE itself.
3897
3898 (ii) The pcre_extra structure (see 45 above) has a void * field called
3899 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3900 pcre_callout_block structure has a field of the same name. The contents of
3901 the field passed in the pcre_extra structure are passed to the callout
3902 function in the corresponding field in the callout block. This makes it
3903 easier to use the same callout-containing regex from multiple threads. For
3904 testing, the pcretest program has a new data escape
3905
3906 \C*n pass the number n (may be negative) as callout_data
3907
3908 If the callout function in pcretest receives a non-zero value as
3909 callout_data, it returns that value.
3910
3911 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3912 there were some redundant $(CFLAGS) in commands that are now specified as
3913 $(LINK), which already includes $(CFLAGS).
3914
3915 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3916 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3917 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3918 one-byte characters throughout. Note that case-insensitive matching applies
3919 only to characters whose values are less than 256. PCRE doesn't support the
3920 notion of cases for higher-valued characters.
3921
3922 (i) A character class whose characters are all within 0-255 is handled as
3923 a bit map, and the map is inverted for negative classes. Previously, a
3924 character > 255 always failed to match such a class; however it should
3925 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3926
3927 (ii) A negated character class with a single character < 255 is coded as
3928 "not this character" (OP_NOT). This wasn't working properly when the test
3929 character was multibyte, either singly or repeated.
3930
3931 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3932 mode, for example: \x{100}{2,3}.
3933
3934 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3935 singly or repeated) now correctly test multibyte characters. However,
3936 PCRE doesn't recognize any characters with values greater than 255 as
3937 digits, spaces, or word characters. Such characters always match \D, \S,
3938 and \W, and never match \d, \s, or \w.
3939
3940 (v) Classes may now contain characters and character ranges with values
3941 greater than 255. For example: [ab\x{100}-\x{400}].
3942
3943 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3944 PCRE in UTF-8 mode.
3945
3946 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3947 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3948 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3949 value.)
3950
3951 53. The single man page has become too large. I have therefore split it up into
3952 a number of separate man pages. These also give rise to individual HTML pages;
3953 these are now put in a separate directory, and there is an index.html page that
3954 lists them all. Some hyperlinking between the pages has been installed.
3955
3956 54. Added convenience functions for handling named capturing parentheses.
3957
3958 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3959 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3960 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3961 are faulted.
3962
3963 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3964 calling configure. These values are used when compiling the dftables.c program
3965 which is run to generate the source of the default character tables. They
3966 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3967 you will need to set these values.
3968
3969 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3970
3971
3972 Version 3.9 02-Jan-02
3973 ---------------------
3974
3975 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3976
3977 2. If --disable-static was given, the building process failed when trying to
3978 build pcretest and pcregrep. (For some reason it was using libtool to compile
3979 them, which is not right, as they aren't part of the library.)
3980
3981
3982 Version 3.8 18-Dec-01
3983 ---------------------
3984
3985 1. The experimental UTF-8 code was completely screwed up. It was packing the
3986 bytes in the wrong order. How dumb can you get?
3987
3988
3989 Version 3.7 29-Oct-01
3990 ---------------------
3991
3992 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3993 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3994 this didn't happen under Solaris 8, where I normally test things.
3995
3996 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3997 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3998 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3999 this fix an hour or so after the initial 3.7 release.)
4000
4001
4002 Version 3.6 23-Oct-01
4003 ---------------------
4004
4005 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4006 offsets passed as NULL with zero offset count.
4007
4008 2. The config.guess and config.sub files had not been updated when I moved to
4009 the latest autoconf.
4010
4011
4012 Version 3.5 15-Aug-01
4013 ---------------------
4014
4015 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4016 had been forgotten.
4017
4018 2. By using declared but undefined structures, we can avoid using "void"
4019 definitions in pcre.h while keeping the internal definitions of the structures
4020 private.
4021
4022 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4023 user point of view, this means that both static and shared libraries are built
4024 by default, but this can be individually controlled. More of the work of
4025 handling this static/shared cases is now inside libtool instead of PCRE's make
4026 file.
4027
4028 4. The pcretest utility is now installed along with pcregrep because it is
4029 useful for users (to test regexs) and by doing this, it automatically gets
4030 relinked by libtool. The documentation has been turned into a man page, so
4031 there are now .1, .txt, and .html versions in /doc.
4032
4033 5. Upgrades to pcregrep:
4034 (i) Added long-form option names like gnu grep.
4035 (ii) Added --help to list all options with an explanatory phrase.
4036 (iii) Added -r, --recursive to recurse into sub-directories.
4037 (iv) Added -f, --file to read patterns from a file.
4038
4039 6. pcre_exec() was referring to its "code" argument before testing that
4040 argument for NULL (and giving an error if it was NULL).
4041
4042 7. Upgraded Makefile.in to allow for compiling in a different directory from
4043 the source directory.
4044
4045 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4046 options bits, the pointer it was passed was to an int instead of to an unsigned
4047 long int. This mattered only on 64-bit systems.
4048
4049 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4050 generated) instead of pcre.in, which it its source. Also made the same change
4051 in several of the .c files.
4052
4053 10. A new release of gcc defines printf() as a macro, which broke pcretest
4054 because it had an ifdef in the middle of a string argument for printf(). Fixed
4055 by using separate calls to printf().
4056
4057 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4058 script, to force use of CR or LF instead of \n in the source. On non-Unix
4059 systems, the value can be set in config.h.
4060
4061 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4062 absolute limit. Changed the text of the error message to make this clear, and
4063 likewise updated the man page.
4064
4065 13. The limit of 99 on the number of capturing subpatterns has been removed.
4066 The new limit is 65535, which I hope will not be a "real" limit.
4067
4068
4069 Version 3.4 22-Aug-00
4070 ---------------------
4071
4072 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4073
4074 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4075
4076
4077 Version 3.3 01-Aug-00
4078 ---------------------
4079
4080 1. If an octal character was given, but the value was greater than \377, it
4081 was not getting masked to the least significant bits, as documented. This could
4082 lead to crashes in some systems.
4083
4084 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4085 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4086
4087 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4088 These just pass their arguments on to (pcre_free)(), but they are provided
4089 because some uses of PCRE bind it to non-C systems that can call its functions,
4090 but cannot call free() or pcre_free() directly.
4091
4092 4. Add "make test" as a synonym for "make check". Corrected some comments in
4093 the Makefile.
4094
4095 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4096 Makefile.
4097
4098 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4099 command called pgrep for grepping around the active processes.
4100
4101 7. Added the beginnings of support for UTF-8 character strings.
4102
4103 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4104 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4105 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4106 out for the ar command.)
4107
4108
4109 Version 3.2 12-May-00
4110 ---------------------
4111
4112 This is purely a bug fixing release.
4113
4114 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4115 of ZA. This was just one example of several cases that could provoke this bug,
4116 which was introduced by change 9 of version 2.00. The code for breaking
4117 infinite loops after an iteration that matches an empty string was't working
4118 correctly.
4119
4120 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4121 when matched against abbab (for example). After matching an empty string, it
4122 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4123 caused it to match further down the string than it should.
4124
4125 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4126 was there because it doesn't seem to be needed, and it causes trouble on some
4127 systems, as it is not a Standard C header. It has been removed.
4128
4129 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4130 were reported on the Macintosh. The changes were from
4131
4132 while ((c = *(++ptr)) != 0 && c != '\n');
4133 to
4134 while ((c = *(++ptr)) != 0 && c != '\n') ;
4135
4136 Totally extraordinary, but if that's what it takes...
4137
4138 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4139 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4140 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4141 assumes the way PCRE uses memmove() (always moving upwards).
4142
4143 6. PCRE is being used in one environment where strchr() is not available. There
4144 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4145 faster code anyway.
4146
4147
4148 Version 3.1 09-Feb-00
4149 ---------------------
4150
4151 The only change in this release is the fixing of some bugs in Makefile.in for
4152 the "install" target:
4153
4154 (1) It was failing to install pcreposix.h.
4155
4156 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4157
4158
4159 Version 3.0 01-Feb-00
4160 ---------------------
4161
4162 1. Add support for the /+ modifier to perltest (to output $` like it does in
4163 pcretest).
4164
4165 2. Add support for the /g modifier to perltest.
4166
4167 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4168 matches null strings.
4169
4170 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4171 pattern. Perl treats empty patterns specially - it reuses the most recent
4172 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4173 effect.
4174
4175 5. The POSIX interface was broken in that it was just handing over the POSIX
4176 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4177 required a bigger vector, with some working space on the end. This means that
4178 the POSIX wrapper now has to get and free some memory, and copy the results.
4179
4180 6. Added some simple autoconf support, placing the test data and the
4181 documentation in separate directories, re-organizing some of the
4182 information files, and making it build pcre-config (a GNU standard). Also added
4183 libtool support for building PCRE as a shared library, which is now the
4184 default.
4185
4186 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4187 09 are not valid octal constants. Single digits will be used for minor values
4188 less than 10.
4189
4190 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4191 existing programs that set these in the POSIX interface can use PCRE without
4192 modification.
4193
4194 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4195 return all that pcre_info() returns, plus additional data. The pcre_info()
4196 function is retained for compatibility, but is considered to be obsolete.
4197
4198 10. Added experimental recursion feature (?R) to handle one common case that
4199 Perl 5.6 will be able to do with (?p{...}).
4200
4201 11. Added support for POSIX character classes like [:alpha:], which Perl is
4202 adopting.
4203
4204
4205 Version 2.08 31-Aug-99
4206 ----------------------
4207
4208 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4209 trying to match at the startoffset position, but instead was moving forward to
4210 the next newline as if a previous match had failed.
4211
4212 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4213 and could get into a loop if a null string was matched other than at the start
4214 of the subject.
4215
4216 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4217 be distinguished at compile time, and for completeness also added PCRE_DATE.
4218
4219 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4220 in GnuWin32 environments.
4221
4222
4223 Version 2.07 29-Jul-99
4224 ----------------------
4225
4226 1. The documentation is now supplied in plain text form and HTML as well as in
4227 the form of man page sources.
4228
4229 2. C++ compilers don't like assigning (void *) values to other pointer types.
4230 In particular this affects malloc(). Although there is no problem in Standard
4231 C, I've put in casts to keep C++ compilers happy.
4232
4233 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4234 should be (const char *).
4235
4236 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4237 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4238 However, I haven't made this a standard facility. The documentation doesn't
4239 mention it, and the Makefile doesn't support it.
4240
4241 5. The Makefile now contains an "install" target, with editable destinations at
4242 the top of the file. The pcretest program is not installed.
4243
4244 6. pgrep -V now gives the PCRE version number and date.
4245
4246 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4247 causing the entire string to be ignored, instead of just the last character.
4248
4249 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4250 non-matching string, it can take a very, very long time, even for strings of
4251 quite modest length, because of the nested recursion. PCRE now does better in
4252 some of these cases. It does this by remembering the last required literal
4253 character in the pattern, and pre-searching the subject to ensure it is present
4254 before running the real match. In other words, it applies a heuristic to detect
4255 some types of certain failure quickly, and in the above example, if presented
4256 with a string that has no trailing " it gives "no match" very quickly.
4257
4258 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4259 other alternatives are tried instead.
4260
4261
4262 Version 2.06 09-Jun-99
4263 ----------------------
4264
4265 1. Change pcretest's output for amount of store used to show just the code
4266 space, because the remainder (the data block) varies in size between 32-bit and
4267 64-bit systems.
4268
4269 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4270 start matching at. This allows lookbehinds to work when searching for multiple
4271 occurrences in a string.
4272
4273 3. Added additional options to pcretest for testing multiple occurrences:
4274
4275 /+ outputs the rest of the string that follows a match
4276 /g loops for multiple occurrences, using the new startoffset argument
4277 /G loops for multiple occurrences by passing an incremented pointer
4278
4279 4. PCRE wasn't doing the "first character" optimization for patterns starting
4280 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4281 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4282 the letter 'x'. On long subject strings, this gives a significant speed-up.
4283
4284
4285 Version 2.05 21-Apr-99
4286 ----------------------
4287
4288 1. Changed the type of magic_number from int to long int so that it works
4289 properly on 16-bit systems.
4290
4291 2. Fixed a bug which caused patterns starting with .* not to work correctly
4292 when the subject string contained newline characters. PCRE was assuming
4293 anchoring for such patterns in all cases, which is not correct because .* will
4294 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4295 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4296 must be retried after every newline in the subject.
4297
4298
4299 Version 2.04 18-Feb-99
4300 ----------------------
4301
4302 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4303 computation of the store needed to hold the pattern was incorrect (too large).
4304 If such patterns were nested a few deep, this could multiply and become a real
4305 problem.
4306
4307 2. Added /M option to pcretest to show the memory requirement of a specific
4308 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4309
4310 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4311 compiled in such a way that the backtracking after subsequent failure was
4312 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4313 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4314
4315
4316 Version 2.03 02-Feb-99
4317 ----------------------
4318
4319 1. Fixed typo and small mistake in man page.
4320
4321 2. Added 4th condition (GPL supersedes if conflict) and created separate
4322 LICENCE file containing the conditions.
4323
4324 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4325 Perl, that is the internal \ allows the delimiter to be included in the
4326 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4327 the final delimiter, add \ to the end of the pattern (to test the error).
4328
4329 4. Added the convenience functions for extracting substrings after a successful
4330 match. Updated pcretest to make it able to test these functions.
4331
4332
4333 Version 2.02 14-Jan-99
4334 ----------------------
4335
4336 1. Initialized the working variables associated with each extraction so that
4337 their saving and restoring doesn't refer to uninitialized store.
4338
4339 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4340 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4341 fix the problem.
4342
4343 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4344 calls, and wasn't printing the correct value for compiling calls. Increased the
4345 default value of LOOPREPEAT, and the number of significant figures in the
4346 times.
4347
4348 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4349
4350 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4351 a building problem on Windows NT with a FAT file system.
4352
4353
4354 Version 2.01 21-Oct-98
4355 ----------------------
4356
4357 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4358 to character tables built by pcre_maketables() in the current locale. If NULL
4359 is passed, the default tables are used.
4360
4361
4362 Version 2.00 24-Sep-98
4363 ----------------------
4364
4365 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4366 it any more.
4367
4368 2. Allow quantification of (?>) groups, and make it work correctly.
4369
4370 3. The first character computation wasn't working for (?>) groups.
4371
4372 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4373 end of the subject) and add 5.005's \z, which really does match only at the
4374 very end of the subject.
4375
4376 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4377
4378 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4379 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4380 localized options. All options to pcre_study() were also removed.
4381
4382 7. Add other new features from 5.005:
4383
4384 $(?<= positive lookbehind
4385 $(?<! negative lookbehind
4386 (?imsx-imsx) added the unsetting capability
4387 such a setting is global if at outer level; local otherwise
4388 (?imsx-imsx:) non-capturing groups with option setting
4389 (?(cond)re|re) conditional pattern matching
4390
4391 A backreference to itself in a repeated group matches the previous
4392 captured string.
4393
4394 8. General tidying up of studying (both automatic and via "study")
4395 consequential on the addition of new assertions.
4396
4397 9. As in 5.005, unlimited repeated groups that could match an empty substring
4398 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4399 runtime if any iteration does actually match an empty substring.
4400
4401 10. Include the RunTest script in the distribution.
4402
4403 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4404 discrepancies, some of which were old and were also with respect to 5.004. They
4405 have now been fixed.
4406
4407
4408 Version 1.09 28-Apr-98
4409 ----------------------
4410
4411 1. A negated single character class followed by a quantifier with a minimum
4412 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4413 program crashes, or just wrong answers. This did not apply to negated classes
4414 containing more than one character, or to minima other than one.
4415
4416
4417 Version 1.08 27-Mar-98
4418 ----------------------
4419
4420 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4421
4422 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4423 latter must appear before anything that relies on it in the pattern.
4424
4425
4426 Version 1.07 16-Feb-98
4427 ----------------------
4428
4429 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4430 repeat of a potentially empty string).
4431
4432
4433 Version 1.06 23-Jan-98
4434 ----------------------
4435
4436 1. Added Markus Oberhumer's little patches for C++.
4437
4438 2. Literal strings longer than 255 characters were broken.
4439
4440
4441 Version 1.05 23-Dec-97
4442 ----------------------
4443
4444 1. Negated character classes containing more than one character were failing if
4445 PCRE_CASELESS was set at run time.
4446
4447
4448 Version 1.04 19-Dec-97
4449 ----------------------
4450
4451 1. Corrected the man page, where some "const" qualifiers had been omitted.
4452
4453 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4454 input syntax.
4455
4456 3. Fixed memory leak which occurred when a regex with back references was
4457 matched with an offsets vector that wasn't big enough. The temporary memory
4458 that is used in this case wasn't being freed if the match failed.
4459
4460 4. Tidied pcretest to ensure it frees memory that it gets.
4461
4462 5. Temporary memory was being obtained in the case where the passed offsets
4463 vector was exactly big enough.
4464
4465 6. Corrected definition of offsetof() from change 5 below.
4466
4467 7. I had screwed up change 6 below and broken the rules for the use of
4468 setjmp(). Now fixed.
4469
4470
4471 Version 1.03 18-Dec-97
4472 ----------------------
4473
4474 1. A erroneous regex with a missing opening parenthesis was correctly
4475 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4476 on some systems.
4477
4478 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4479 it was reported that one broken compiler failed on the former because "code" is
4480 also an independent variable.
4481
4482 3. The erroneous regex a[]b caused an array overrun reference.
4483
4484 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4485 fail on data ending with that character. (It was going on too far, and checking
4486 the next character, typically a binary zero.) This was specific to the
4487 optimized code for single-character negative classes.
4488
4489 5. Added a contributed patch from the TIN world which does the following:
4490
4491 + Add an undef for memmove, in case the the system defines a macro for it.
4492
4493 + Add a definition of offsetof(), in case there isn't one. (I don't know
4494 the reason behind this - offsetof() is part of the ANSI standard - but
4495 it does no harm).
4496
4497 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4498 most of the places where whitespace preceded '#'. I have given up and
4499 allowed the remaining 2 cases to be at the margin.
4500
4501 + Rename some variables in pcre to eliminate shadowing. This seems very
4502 pedantic, but does no harm, of course.
4503
4504 6. Moved the call to setjmp() into its own function, to get rid of warnings
4505 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4506
4507 7. Constructs such as \d{8,} were compiling into the equivalent of
4508 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4509 outcome, but in this particular case used more store than had been allocated,
4510 which caused the bug to be discovered because it threw up an internal error.
4511
4512 8. The debugging code in both pcre and pcretest for outputting the compiled
4513 form of a regex was going wrong in the case of back references followed by
4514 curly-bracketed repeats.
4515
4516
4517 Version 1.02 12-Dec-97
4518 ----------------------
4519
4520 1. Typos in pcre.3 and comments in the source fixed.
4521
4522 2. Applied a contributed patch to get rid of places where it used to remove
4523 'const' from variables, and fixed some signed/unsigned and uninitialized
4524 variable warnings.
4525
4526 3. Added the "runtest" target to Makefile.
4527
4528 4. Set default compiler flag to -O2 rather than just -O.
4529
4530
4531 Version 1.01 19-Nov-97
4532 ----------------------
4533
4534 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4535 like /([ab]*)*/, that is, for classes with more than one character in them.
4536
4537 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4538 as /((?>a*))*/ (a PCRE_EXTRA facility).
4539
4540
4541 Version 1.00 18-Nov-97
4542 ----------------------
4543
4544 1. Added compile-time macros to support systems such as SunOS4 which don't have
4545 memmove() or strerror() but have other things that can be used instead.
4546
4547 2. Arranged that "make clean" removes the executables.
4548
4549
4550 Version 0.99 27-Oct-97
4551 ----------------------
4552
4553 1. Fixed bug in code for optimizing classes with only one character. It was
4554 initializing a 32-byte map regardless, which could cause it to run off the end
4555 of the memory it had got.
4556
4557 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4558
4559
4560 Version 0.98 22-Oct-97
4561 ----------------------
4562
4563 1. Fixed bug in code for handling temporary memory usage when there are more
4564 back references than supplied space in the ovector. This could cause segfaults.
4565
4566
4567 Version 0.97 21-Oct-97
4568 ----------------------
4569
4570 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4571
4572 2. Optimized negated single characters not to use a bit map.
4573
4574 3. Brought error texts together as macro definitions; clarified some of them;
4575 fixed one that was wrong - it said "range out of order" when it meant "invalid
4576 escape sequence".
4577
4578 4. Changed some char * arguments to const char *.
4579
4580 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4581
4582 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4583 pcretest.
4584
4585
4586 Version 0.96 16-Oct-97
4587 ----------------------
4588
4589 1. Added a simple "pgrep" utility to the distribution.
4590
4591 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4592 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4593 where "ddd" means "one or more decimal digits".
4594
4595 3. Fixed serious bug. If a pattern had a back reference, but the call to
4596 pcre_exec() didn't supply a large enough ovector to record the related
4597 identifying subpattern, the match always failed. PCRE now remembers the number
4598 of the largest back reference, and gets some temporary memory in which to save
4599 the offsets during matching if necessary, in order to ensure that
4600 backreferences always work.
4601
4602 4. Increased the compatibility with Perl in a number of ways:
4603
4604 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4605 to request this handling. The option can be set at compile or exec time.
4606
4607 (b) $ matches before a terminating newline by default; an option
4608 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4609 mode). The option can be set at compile or exec time.
4610
4611 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4612 the same as Perl's. If the decimal number it represents is less than 10
4613 or there aren't that many previous left capturing parentheses, an octal
4614 escape is read. Inside a character class, it's always an octal escape,
4615 even if it is a single digit.
4616
4617 (d) An escaped but undefined alphabetic character is taken as a literal,
4618 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4619 escapes.
4620
4621 (e) {0} is now permitted. (The previous item is removed from the compiled
4622 pattern).
4623
4624 5. Changed all the names of code files so that the basic parts are no longer
4625 than 10 characters, and abolished the teeny "globals.c" file.
4626
4627 6. Changed the handling of character classes; they are now done with a 32-byte
4628 bit map always.
4629
4630 7. Added the -d and /D options to pcretest to make it possible to look at the
4631 internals of compilation without having to recompile pcre.
4632
4633
4634 Version 0.95 23-Sep-97
4635 ----------------------
4636
4637 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4638 \x20 at the start of a run of normal characters. These were being treated as
4639 real characters, instead of the source characters being re-checked.
4640
4641
4642 Version 0.94 18-Sep-97
4643 ----------------------
4644
4645 1. The functions are now thread-safe, with the caveat that the global variables
4646 containing pointers to malloc() and free() or alternative functions are the
4647 same for all threads.
4648
4649 2. Get pcre_study() to generate a bitmap of initial characters for non-
4650 anchored patterns when this is possible, and use it if passed to pcre_exec().
4651
4652
4653 Version 0.93 15-Sep-97
4654 ----------------------
4655
4656 1. /(b)|(:+)/ was computing an incorrect first character.
4657
4658 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4659 but not actually doing anything yet.
4660
4661 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4662 as Perl does (e.g. [-az] or [az-]).
4663
4664 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4665 all possible positions.
4666
4667 5. Split up into different modules to avoid including unneeded functions in a
4668 compiled binary. However, compile and exec are still in one module. The "study"
4669 function is split off.
4670
4671 6. The character tables are now in a separate module whose source is generated
4672 by an auxiliary program - but can then be edited by hand if required. There are
4673 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4674 toupper() in the code.
4675
4676 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4677 make them global. Abolish the function for setting them, as the caller can now
4678 set them directly.
4679
4680
4681 Version 0.92 11-Sep-97
4682 ----------------------
4683
4684 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4685 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4686
4687 2. Caseless matching was not working in character classes if the characters in
4688 the pattern were in upper case.
4689
4690 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4691
4692 4. Make PCRE_ANCHORED public and accept as a compile option.
4693
4694 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4695 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4696 pass them.
4697
4698 6. Give an error if bad option bits passed at compile or run time.
4699
4700 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4701 pcretest to cause it to pass that flag.
4702
4703 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4704 options, and the first character, if set.
4705
4706 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4707
4708
4709 Version 0.91 10-Sep-97
4710 ----------------------
4711
4712 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4713 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4714
4715 2. PCRE was looping on encountering an indefinitely repeated back reference to
4716 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4717 Perl does - treats the match as successful.
4718
4719 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5