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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 932 - (show annotations) (download)
Fri Feb 24 18:54:43 2012 UTC (2 years, 5 months ago) by ph10
File size: 227762 byte(s)
Add support for PCRE_INFO_MAXLOOKBEHIND.

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

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12