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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 942 - (show annotations)
Tue Feb 28 14:50:31 2012 UTC (3 years, 5 months ago) by ph10
File size: 228536 byte(s)
Error occurred while calculating annotation data.
Update for Unicode 6.1.0.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.31
5 -----------------------------
6
7 1. Fixing a wrong JIT test case and some compiler warnings.
8
9 2. Removed a bashism from the RunTest script.
10
11 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
12 to unsigned type, result still unsigned" that was given by an MS compiler
13 on encountering the code "-sizeof(xxx)".
14
15 4. Partial matching support is added to the JIT compiler.
16
17 5. Fixed several bugs concerned with partial matching of items that consist
18 of more than one character:
19
20 (a) /^(..)\1/ did not partially match "aba" because checking references was
21 done on an "all or nothing" basis. This also applied to repeated
22 references.
23
24 (b) \R did not give a hard partial match if \r was found at the end of the
25 subject.
26
27 (c) \X did not give a hard partial match after matching one or more
28 characters at the end of the subject.
29
30 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
31 a partial match for the string "\r".
32
33 (e) When newline was set to CRLF, the metacharacter "." did not recognize
34 a partial match for a CR character at the end of the subject string.
35
36 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
37 running pcretest, the text "(JIT)" added to the output whenever JIT is
38 actually used to run the match.
39
40 7. Individual JIT compile options can be set in pcretest by following -s+[+]
41 or /S+[+] with a digit between 1 and 7.
42
43 8. OP_NOT now supports any UTF character not just single-byte ones.
44
45 9. (*MARK) control verb is now supported by the JIT compiler.
46
47 10. The command "./RunTest list" lists the available tests without actually
48 running any of them. (Because I keep forgetting what they all are.)
49
50 11. Add PCRE_INFO_MAXLOOKBEHIND.
51
52 12. Applied a (slightly modified) user-supplied patch that improves performance
53 when the heap is used for recursion (compiled with --disable-stack-for-
54 recursion). Instead of malloc and free for each heap frame each time a
55 logical recursion happens, frames are retained on a chain and re-used where
56 possible. This sometimes gives as much as 30% improvement.
57
58 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
59 call.
60
61 14. As documented, (*COMMIT) is now confined to within a positive assertion.
62
63 15. It is now possible to link pcretest with libedit as an alternative to
64 libreadline.
65
66 16. (*COMMIT) control verb is now supported by the JIT compiler.
67
68 6. The Unicode data tables have been updated to Unicode 6.1.0.
69
70
71 Version 8.30 04-February-2012
72 -----------------------------
73
74 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
75 name is defined in ctype.h.
76
77 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
78 only in quite long subpatterns.
79
80 3. Removed the function pcre_info(), which has been obsolete and deprecated
81 since it was replaced by pcre_fullinfo() in February 2000.
82
83 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
84 match a (*MARK), and the match failed at the start of the subject, a
85 reference to memory before the start of the subject could occur. This bug
86 was introduced by fix 17 of release 8.21.
87
88 5. A reference to an unset group with zero minimum repetition was giving
89 totally wrong answers (in non-JavaScript-compatibility mode). For example,
90 /(another)?(\1?)test/ matched against "hello world test". This bug was
91 introduced in release 8.13.
92
93 6. Add support for 16-bit character strings (a large amount of work involving
94 many changes and refactorings).
95
96 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
97 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
98 from a file.
99
100 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
101 rounding is not applied in this particular case).
102
103 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
104 if they appear, or are escaped, in patterns.
105
106 10. Get rid of a number of -Wunused-but-set-variable warnings.
107
108 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
109 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
110 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
111 also returns the mark "x". This bug applied to capturing parentheses,
112 non-capturing parentheses, and atomic parentheses. It also applied to some
113 assertions.
114
115 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
116 information out of configure.ac instead of relying on pcre.h.generic, which
117 is not stored in the repository.
118
119 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
120 -lreadline.
121
122 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
123
124 15. Applied Graycode's patch to put the top-level frame on the stack rather
125 than the heap when not using the stack for recursion. This gives a
126 performance improvement in many cases when recursion is not deep.
127
128 16. Experimental code added to "pcretest -C" to output the stack frame size.
129
130
131 Version 8.21 12-Dec-2011
132 ------------------------
133
134 1. Updating the JIT compiler.
135
136 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
137 are added as well.
138
139 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
140 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
141 calling _pcre_jit_exec. Some extra comments are added.
142
143 4. (*MARK) settings inside atomic groups that do not contain any capturing
144 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
145 was introduced by change 18 for 8.20.
146
147 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
148 ECMA-262 standard.
149
150 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
151 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
152 This bug was probably introduced by change 9 of 8.13.
153
154 7. While fixing 6 above, I noticed that a number of other items were being
155 incorrectly rejected as "not fixed length". This arose partly because newer
156 opcodes had not been added to the fixed-length checking code. I have (a)
157 corrected the bug and added tests for these items, and (b) arranged for an
158 error to occur if an unknown opcode is encountered while checking for fixed
159 length instead of just assuming "not fixed length". The items that were
160 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
161 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
162 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
163
164 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
165 being incorrectly compiled and would have given unpredicatble results.
166
167 9. A possessively repeated subpattern with minimum repeat count greater than
168 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
169 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
170 the first (A) could occur when it should not.
171
172 10. Add a cast and remove a redundant test from the code.
173
174 11. JIT should use pcre_malloc/pcre_free for allocation.
175
176 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
177 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
178 is anything other than /usr, -L is still shown).
179
180 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
181
182 14. Perl does not support \N without a following name in a [] class; PCRE now
183 also gives an error.
184
185 15. If a forward reference was repeated with an upper limit of around 2000,
186 it caused the error "internal error: overran compiling workspace". The
187 maximum number of forward references (including repeats) was limited by the
188 internal workspace, and dependent on the LINK_SIZE. The code has been
189 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
190 the default depends on LINK_SIZE. There is a new upper limit (for safety)
191 of around 200,000 forward references. While doing this, I also speeded up
192 the filling in of repeated forward references.
193
194 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
195 incorrectly expecting the subject to contain another "a" after the start.
196
197 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
198 in the match, the SKIP should be ignored. This was not happening; instead
199 the SKIP was being treated as NOMATCH. For patterns such as
200 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
201 tested.
202
203 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
204 now much more compatible with Perl, in particular in cases where the result
205 is a non-match for a non-anchored pattern. For example, if
206 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
207 "m", where previously it did not return a name. A side effect of this
208 change is that for partial matches, the last encountered mark name is
209 returned, as for non matches. A number of tests that were previously not
210 Perl-compatible have been moved into the Perl-compatible test files. The
211 refactoring has had the pleasing side effect of removing one argument from
212 the match() function, thus reducing its stack requirements.
213
214 19. If the /S+ option was used in pcretest to study a pattern using JIT,
215 subsequent uses of /S (without +) incorrectly behaved like /S+.
216
217 21. Retrieve executable code size support for the JIT compiler and fixing
218 some warnings.
219
220 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
221 not work when the shorter character appeared right at the end of the
222 subject string.
223
224 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
225 systems.
226
227 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
228 output it when the /M option is used in pcretest.
229
230 25. The CheckMan script was not being included in the distribution. Also, added
231 an explicit "perl" to run Perl scripts from the PrepareRelease script
232 because this is reportedly needed in Windows.
233
234 26. If study data was being save in a file and studying had not found a set of
235 "starts with" bytes for the pattern, the data written to the file (though
236 never used) was taken from uninitialized memory and so caused valgrind to
237 complain.
238
239 27. Updated RunTest.bat as provided by Sheri Pierce.
240
241 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
242
243 29. Computation of memory usage for the table of capturing group names was
244 giving an unnecessarily large value.
245
246
247 Version 8.20 21-Oct-2011
248 ------------------------
249
250 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
251 a POSIX class. After further experiments with Perl, which convinced me that
252 Perl has bugs and confusions, a closing square bracket is no longer allowed
253 in a POSIX name. This bug also affected patterns with classes that started
254 with full stops.
255
256 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
257 captured substring, but while checking the failing first alternative,
258 substring 1 is temporarily captured. If the output vector supplied to
259 pcre_exec() was not big enough for this capture, the yield of the function
260 was still zero ("insufficient space for captured substrings"). This cannot
261 be totally fixed without adding another stack variable, which seems a lot
262 of expense for a edge case. However, I have improved the situation in cases
263 such as /(a)(b)x|abc/ matched against "abc", where the return code
264 indicates that fewer than the maximum number of slots in the ovector have
265 been set.
266
267 3. Related to (2) above: when there are more back references in a pattern than
268 slots in the output vector, pcre_exec() uses temporary memory during
269 matching, and copies in the captures as far as possible afterwards. It was
270 using the entire output vector, but this conflicts with the specification
271 that only 2/3 is used for passing back captured substrings. Now it uses
272 only the first 2/3, for compatibility. This is, of course, another edge
273 case.
274
275 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
276 main code base, and can be used by building with --enable-jit. When this is
277 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
278 runtime --no-jit option is given.
279
280 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
281 ovector, the return from the function was zero, implying that there were
282 other matches that did not fit. The correct "exactly full" value is now
283 returned.
284
285 6. If a subpattern that was called recursively or as a subroutine contained
286 (*PRUNE) or any other control that caused it to give a non-standard return,
287 invalid errors such as "Error -26 (nested recursion at the same subject
288 position)" or even infinite loops could occur.
289
290 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
291 computing the minimum length on reaching *ACCEPT, and so ended up with the
292 wrong value of 1 rather than 0. Further investigation indicates that
293 computing a minimum subject length in the presence of *ACCEPT is difficult
294 (think back references, subroutine calls), and so I have changed the code
295 so that no minimum is registered for a pattern that contains *ACCEPT.
296
297 8. If (*THEN) was present in the first (true) branch of a conditional group,
298 it was not handled as intended. [But see 16 below.]
299
300 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
301 Sheri Pierce.
302
303 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
304 the first byte in a match must be "a".
305
306 11. Change 17 for 8.13 increased the recursion depth for patterns like
307 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
308 pattern contains any instances of (*THEN). If it does not, the old
309 optimizations are restored. It would be nice to do this on a per-group
310 basis, but at the moment that is not feasible.
311
312 12. In some environments, the output of pcretest -C is CRLF terminated. This
313 broke RunTest's code that checks for the link size. A single white space
314 character after the value is now allowed for.
315
316 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
317 For "fr", it uses the Windows-specific input and output files.
318
319 14. If (*THEN) appeared in a group that was called recursively or as a
320 subroutine, it did not work as intended. [But see next item.]
321
322 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
323 pattern fragments (but not containing any | characters). If A and B are
324 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
325 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
326 D. In other words, Perl considers parentheses that do not contain any |
327 characters to be part of a surrounding alternative, whereas PCRE was
328 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
329 differently. PCRE now behaves in the same way as Perl, except in the case
330 of subroutine/recursion calls such as (?1) which have in any case always
331 been different (but PCRE had them first :-).
332
333 16. Related to 15 above: Perl does not treat the | in a conditional group as
334 creating alternatives. Such a group is treated in the same way as an
335 ordinary group without any | characters when processing (*THEN). PCRE has
336 been changed to match Perl's behaviour.
337
338 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
339 RunGrepTest script failed.
340
341 18. Change 22 for version 13 caused atomic groups to use more stack. This is
342 inevitable for groups that contain captures, but it can lead to a lot of
343 stack use in large patterns. The old behaviour has been restored for atomic
344 groups that do not contain any capturing parentheses.
345
346 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
347 suppress the check for a minimum subject length at run time. (If it was
348 given to pcre_exec() or pcre_dfa_exec() it did work.)
349
350 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
351 fail to work when decoding hex characters in data strings in EBCDIC
352 environments.
353
354 21. It appears that in at least one Mac OS environment, the isxdigit() function
355 is implemented as a macro that evaluates to its argument more than once,
356 contravening the C 90 Standard (I haven't checked a later standard). There
357 was an instance in pcretest which caused it to go wrong when processing
358 \x{...} escapes in subject strings. The has been rewritten to avoid using
359 things like p++ in the argument of isxdigit().
360
361
362 Version 8.13 16-Aug-2011
363 ------------------------
364
365 1. The Unicode data tables have been updated to Unicode 6.0.0.
366
367 2. Two minor typos in pcre_internal.h have been fixed.
368
369 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
370 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
371 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
372
373 4. There were a number of related bugs in the code for matching backrefences
374 caselessly in UTF-8 mode when codes for the characters concerned were
375 different numbers of bytes. For example, U+023A and U+2C65 are an upper
376 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
377 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
378 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
379 2-byte code at the end of the subject (it thought there wasn't enough data
380 left).
381
382 5. Comprehensive information about what went wrong is now returned by
383 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
384 as the output vector has at least 2 elements. The offset of the start of
385 the failing character and a reason code are placed in the vector.
386
387 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
388 now returned is for the first byte of the failing character, instead of the
389 last byte inspected. This is an incompatible change, but I hope it is small
390 enough not to be a problem. It makes the returned offset consistent with
391 pcre_exec() and pcre_dfa_exec().
392
393 7. pcretest now gives a text phrase as well as the error number when
394 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
395 failure, the offset and reason code are output.
396
397 8. When \R was used with a maximizing quantifier it failed to skip backwards
398 over a \r\n pair if the subsequent match failed. Instead, it just skipped
399 back over a single character (\n). This seems wrong (because it treated the
400 two characters as a single entity when going forwards), conflicts with the
401 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
402 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
403 has been changed.
404
405 9. Some internal refactoring has changed the processing so that the handling
406 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
407 time (the PCRE_DOTALL option was changed this way some time ago: version
408 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
409 which was always a bit of a fudge. It also means that there is one less
410 argument for the match() function, which reduces its stack requirements
411 slightly. This change also fixes an incompatibility with Perl: the pattern
412 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
413
414 10. More internal refactoring has drastically reduced the number of recursive
415 calls to match() for possessively repeated groups such as (abc)++ when
416 using pcre_exec().
417
418 11. While implementing 10, a number of bugs in the handling of groups were
419 discovered and fixed:
420
421 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
422 (a|)*(?1) gave a compile-time internal error.
423 ((a|)+)+ did not notice that the outer group could match an empty string.
424 (^a|^)+ was not marked as anchored.
425 (.*a|.*)+ was not marked as matching at start or after a newline.
426
427 12. Yet more internal refactoring has removed another argument from the match()
428 function. Special calls to this function are now indicated by setting a
429 value in a variable in the "match data" data block.
430
431 13. Be more explicit in pcre_study() instead of relying on "default" for
432 opcodes that mean there is no starting character; this means that when new
433 ones are added and accidentally left out of pcre_study(), testing should
434 pick them up.
435
436 14. The -s option of pcretest has been documented for ages as being an old
437 synonym of -m (show memory usage). I have changed it to mean "force study
438 for every regex", that is, assume /S for every regex. This is similar to -i
439 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
440 using it. It makes it easier to run collections of tests with and without
441 study enabled, and thereby test pcre_study() more easily. All the standard
442 tests are now run with and without -s (but some patterns can be marked as
443 "never study" - see 20 below).
444
445 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
446 restoration of the capturing data to the outer values was not happening
447 correctly.
448
449 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
450 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
451 pattern had matched an empty string, and so incorrectly returned a no
452 match.
453
454 17. There was optimizing code for the last branch of non-capturing parentheses,
455 and also for the obeyed branch of a conditional subexpression, which used
456 tail recursion to cut down on stack usage. Unfortunately, now that there is
457 the possibility of (*THEN) occurring in these branches, tail recursion is
458 no longer possible because the return has to be checked for (*THEN). These
459 two optimizations have therefore been removed. [But see 8.20/11 above.]
460
461 18. If a pattern containing \R was studied, it was assumed that \R always
462 matched two bytes, thus causing the minimum subject length to be
463 incorrectly computed because \R can also match just one byte.
464
465 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
466 was incorrectly computed.
467
468 20. If /S is present twice on a test pattern in pcretest input, it now
469 *disables* studying, thereby overriding the use of -s on the command line
470 (see 14 above). This is necessary for one or two tests to keep the output
471 identical in both cases.
472
473 21. When (*ACCEPT) was used in an assertion that matched an empty string and
474 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
475
476 22. When an atomic group that contained a capturing parenthesis was
477 successfully matched, but the branch in which it appeared failed, the
478 capturing was not being forgotten if a higher numbered group was later
479 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
480 group 1 to "a", when in fact it should be unset. This applied to multi-
481 branched capturing and non-capturing groups, repeated or not, and also to
482 positive assertions (capturing in negative assertions does not happen
483 in PCRE) and also to nested atomic groups.
484
485 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
486 subject after a captured substring, to make it easier to tell which of a
487 number of identical substrings has been captured.
488
489 24. The way atomic groups are processed by pcre_exec() has been changed so that
490 if they are repeated, backtracking one repetition now resets captured
491 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
492 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
493 "aaa". Previously, it would have been "a". As part of this code
494 refactoring, the way recursive calls are handled has also been changed.
495
496 25. If an assertion condition captured any substrings, they were not passed
497 back unless some other capturing happened later. For example, if
498 (?(?=(a))a) was matched against "a", no capturing was returned.
499
500 26. When studying a pattern that contained subroutine calls or assertions,
501 the code for finding the minimum length of a possible match was handling
502 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
503 group 1 called group 2 while simultaneously a separate group 2 called group
504 1). A stack overflow occurred in this case. I have fixed this by limiting
505 the recursion depth to 10.
506
507 27. Updated RunTest.bat in the distribution to the version supplied by Tom
508 Fortmann. This supports explicit test numbers on the command line, and has
509 argument validation and error reporting.
510
511 28. An instance of \X with an unlimited repeat could fail if at any point the
512 first character it looked at was a mark character.
513
514 29. Some minor code refactoring concerning Unicode properties and scripts
515 should reduce the stack requirement of match() slightly.
516
517 30. Added the '=' option to pcretest to check the setting of unused capturing
518 slots at the end of the pattern, which are documented as being -1, but are
519 not included in the return count.
520
521 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
522 compiled something random. Now it gives a compile-time error (as does
523 Perl).
524
525 32. A *MARK encountered during the processing of a positive assertion is now
526 recorded and passed back (compatible with Perl).
527
528 33. If --only-matching or --colour was set on a pcregrep call whose pattern
529 had alternative anchored branches, the search for a second match in a line
530 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
531 matched the line "0102" twice. The same bug affected patterns that started
532 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
533 twice.
534
535 34. Previously, PCRE did not allow quantification of assertions. However, Perl
536 does, and because of capturing effects, quantifying parenthesized
537 assertions may at times be useful. Quantifiers are now allowed for
538 parenthesized assertions.
539
540 35. A minor code tidy in pcre_compile() when checking options for \R usage.
541
542 36. \g was being checked for fancy things in a character class, when it should
543 just be a literal "g".
544
545 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
546 appearance of a nested POSIX class supersedes an apparent external class.
547 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
548 unescaped square brackets may also appear as part of class names. For
549 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
550 more like Perl. (But see 8.20/1 above.)
551
552 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
553 was because it thought it was \N{name}, which is not supported).
554
555 39. Add minix to OS list not supporting the -S option in pcretest.
556
557 40. PCRE tries to detect cases of infinite recursion at compile time, but it
558 cannot analyze patterns in sufficient detail to catch mutual recursions
559 such as ((?1))((?2)). There is now a runtime test that gives an error if a
560 subgroup is called recursively as a subpattern for a second time at the
561 same position in the subject string. In previous releases this might have
562 been caught by the recursion limit, or it might have run out of stack.
563
564 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
565 happen only once. PCRE was, however incorrectly giving a compile time error
566 "recursive call could loop indefinitely" because it cannot analyze the
567 pattern in sufficient detail. The compile time test no longer happens when
568 PCRE is compiling a conditional subpattern, but actual runaway loops are
569 now caught at runtime (see 40 above).
570
571 42. It seems that Perl allows any characters other than a closing parenthesis
572 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
573 has been changed to be the same.
574
575 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
576 as not to get warnings when autogen.sh is called. Also changed
577 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
578
579 44. To help people who use pcregrep to scan files containing exceedingly long
580 lines, the following changes have been made:
581
582 (a) The default value of the buffer size parameter has been increased from
583 8K to 20K. (The actual buffer used is three times this size.)
584
585 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
586 PCRE is built.
587
588 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
589 to be set at run time.
590
591 (d) Numerical values in pcregrep options can be followed by K or M, for
592 example --buffer-size=50K.
593
594 (e) If a line being scanned overflows pcregrep's buffer, an error is now
595 given and the return code is set to 2.
596
597 45. Add a pointer to the latest mark to the callout data block.
598
599 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
600 partial match of an empty string instead of no match. This was specific to
601 the use of ".".
602
603 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
604 complete match instead of a partial match. This bug was dependent on both
605 the PCRE_UTF8 and PCRE_DOTALL options being set.
606
607 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
608 starting byte set, because \b was not being ignored.
609
610
611 Version 8.12 15-Jan-2011
612 ------------------------
613
614 1. Fixed some typos in the markup of the man pages, and wrote a script that
615 checks for such things as part of the documentation building process.
616
617 2. On a big-endian 64-bit system, pcregrep did not correctly process the
618 --match-limit and --recursion-limit options (added for 8.11). In
619 particular, this made one of the standard tests fail. (The integer value
620 went into the wrong half of a long int.)
621
622 3. If the --colour option was given to pcregrep with -v (invert match), it
623 did strange things, either producing crazy output, or crashing. It should,
624 of course, ignore a request for colour when reporting lines that do not
625 match.
626
627 4. Another pcregrep bug caused similar problems if --colour was specified with
628 -M (multiline) and the pattern match finished with a line ending.
629
630 5. In pcregrep, when a pattern that ended with a literal newline sequence was
631 matched in multiline mode, the following line was shown as part of the
632 match. This seems wrong, so I have changed it.
633
634 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
635 the check for further matches in the same line (so they could be coloured)
636 to overrun the end of the current line. If another match was found, it was
637 incorrectly shown (and then shown again when found in the next line).
638
639 7. If pcregrep was compiled under Windows, there was a reference to the
640 function pcregrep_exit() before it was defined. I am assuming this was
641 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
642 reported by a user. I've moved the definition above the reference.
643
644
645 Version 8.11 10-Dec-2010
646 ------------------------
647
648 1. (*THEN) was not working properly if there were untried alternatives prior
649 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
650 backtracked to try for "b" instead of moving to the next alternative branch
651 at the same level (in this case, to look for "c"). The Perl documentation
652 is clear that when (*THEN) is backtracked onto, it goes to the "next
653 alternative in the innermost enclosing group".
654
655 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
656 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
657 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
658 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
659 (*THEN).
660
661 3. If \s appeared in a character class, it removed the VT character from
662 the class, even if it had been included by some previous item, for example
663 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
664 of \s, but is part of the POSIX "space" class.)
665
666 4. A partial match never returns an empty string (because you can always
667 match an empty string at the end of the subject); however the checking for
668 an empty string was starting at the "start of match" point. This has been
669 changed to the "earliest inspected character" point, because the returned
670 data for a partial match starts at this character. This means that, for
671 example, /(?<=abc)def/ gives a partial match for the subject "abc"
672 (previously it gave "no match").
673
674 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
675 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
676 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
677 has an implication that the given string is incomplete (because a partial
678 match is preferred over a full match). For this reason, these items now
679 give a partial match in this situation. [Aside: previously, the one case
680 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
681 match rather than a full match, which was wrong by the old rules, but is
682 now correct.]
683
684 6. There was a bug in the handling of #-introduced comments, recognized when
685 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
686 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
687 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
688 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
689 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
690 places in pcre_compile().
691
692 7. Related to (6) above, when pcre_compile() was skipping #-introduced
693 comments when looking ahead for named forward references to subpatterns,
694 the only newline sequence it recognized was NL. It now handles newlines
695 according to the set newline convention.
696
697 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
698 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
699 cater for a lack of strerror(). These oversights have been fixed.
700
701 9. Added --match-limit and --recursion-limit to pcregrep.
702
703 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
704
705 11. When the -o option was used, pcregrep was setting a return code of 1, even
706 when matches were found, and --line-buffered was not being honoured.
707
708 12. Added an optional parentheses number to the -o and --only-matching options
709 of pcregrep.
710
711 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
712 can match an empty string. The code to do it in pcretest and pcredemo
713 needed fixing:
714
715 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
716 only one byte after an empty string match just before CRLF (this case
717 just got forgotten; "any" and "anycrlf" were OK).
718
719 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
720 mode when an empty string match preceded an ASCII character followed by
721 a non-ASCII character. (The code for advancing by one character rather
722 than one byte was nonsense.)
723
724 (c) The pcredemo.c sample program did not have any code at all to handle
725 the cases when CRLF is a valid newline sequence.
726
727 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
728 as a starting offset was within the subject string. There is now a new
729 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
730 negative or greater than the length of the string. In order to test this,
731 pcretest is extended to allow the setting of negative starting offsets.
732
733 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
734 starting offset points to the beginning of a UTF-8 character was
735 unnecessarily clumsy. I tidied it up.
736
737 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
738 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
739
740 17. Nobody had reported that the --include_dir option, which was added in
741 release 7.7 should have been called --include-dir (hyphen, not underscore)
742 for compatibility with GNU grep. I have changed it to --include-dir, but
743 left --include_dir as an undocumented synonym, and the same for
744 --exclude-dir, though that is not available in GNU grep, at least as of
745 release 2.5.4.
746
747 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
748 characters from a string of bytes) have been redefined so as not to use
749 loops, in order to improve performance in some environments. At the same
750 time, I abstracted some of the common code into auxiliary macros to save
751 repetition (this should not affect the compiled code).
752
753 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
754 compile-time error is now given if \c is not followed by an ASCII
755 character, that is, a byte less than 128. (In EBCDIC mode, the code is
756 different, and any byte value is allowed.)
757
758 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
759 START_OPTIMIZE option, which is now allowed at compile time - but just
760 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
761 to pcregrep and other applications that have no direct access to PCRE
762 options. The new /Y option in pcretest sets this option when calling
763 pcre_compile().
764
765 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
766 back references. Groups containing recursive back references were forced to
767 be atomic by that change, but in the case of named groups, the amount of
768 memory required was incorrectly computed, leading to "Failed: internal
769 error: code overflow". This has been fixed.
770
771 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
772 pcretest.c, to avoid build problems in some Borland environments.
773
774
775 Version 8.10 25-Jun-2010
776 ------------------------
777
778 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
779 THEN.
780
781 2. (*ACCEPT) was not working when inside an atomic group.
782
783 3. Inside a character class, \B is treated as a literal by default, but
784 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
785 causes the error). The code is unchanged, but I tidied the documentation.
786
787 4. Inside a character class, PCRE always treated \R and \X as literals,
788 whereas Perl faults them if its -w option is set. I have changed PCRE so
789 that it faults them when PCRE_EXTRA is set.
790
791 5. Added support for \N, which always matches any character other than
792 newline. (It is the same as "." when PCRE_DOTALL is not set.)
793
794 6. When compiling pcregrep with newer versions of gcc which may have
795 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
796 declared with attribute warn_unused_result" were given. Just casting the
797 result to (void) does not stop the warnings; a more elaborate fudge is
798 needed. I've used a macro to implement this.
799
800 7. Minor change to pcretest.c to avoid a compiler warning.
801
802 8. Added four artifical Unicode properties to help with an option to make
803 \s etc use properties (see next item). The new properties are: Xan
804 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
805
806 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
807 use Unicode properties. (*UCP) at the start of a pattern can be used to set
808 this option. Modified pcretest to add /W to test this facility. Added
809 REG_UCP to make it available via the POSIX interface.
810
811 10. Added --line-buffered to pcregrep.
812
813 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
814 studied, and the match started with a letter with a code point greater than
815 127 whose first byte was different to the first byte of the other case of
816 the letter, the other case of this starting letter was not recognized
817 (#976).
818
819 12. If a pattern that was studied started with a repeated Unicode property
820 test, for example, \p{Nd}+, there was the theoretical possibility of
821 setting up an incorrect bitmap of starting bytes, but fortunately it could
822 not have actually happened in practice until change 8 above was made (it
823 added property types that matched character-matching opcodes).
824
825 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
826 possible starting bytes for non-anchored patterns.
827
828 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
829 \R, and also a number of cases that involve Unicode properties, both
830 explicit and implicit when PCRE_UCP is set.
831
832 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
833 input, it could crash or give wrong results if characters with values
834 greater than 0xc0 were present in the subject string. (Detail: it assumed
835 UTF-8 input when processing these items.)
836
837 16. Added a lot of (int) casts to avoid compiler warnings in systems where
838 size_t is 64-bit (#991).
839
840 17. Added a check for running out of memory when PCRE is compiled with
841 --disable-stack-for-recursion (#990).
842
843 18. If the last data line in a file for pcretest does not have a newline on
844 the end, a newline was missing in the output.
845
846 19. The default pcre_chartables.c file recognizes only ASCII characters (values
847 less than 128) in its various bitmaps. However, there is a facility for
848 generating tables according to the current locale when PCRE is compiled. It
849 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
850 characters, are recognized by isspace() and therefore were getting set in
851 these tables, and indeed these tables seem to approximate to ISO 8859. This
852 caused a problem in UTF-8 mode when pcre_study() was used to create a list
853 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
854 which of course cannot start UTF-8 characters. I have changed the code so
855 that only real ASCII characters (less than 128) and the correct starting
856 bytes for UTF-8 encodings are set for characters greater than 127 when in
857 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
858 altogether.)
859
860 20. Added the /T option to pcretest so as to be able to run tests with non-
861 standard character tables, thus making it possible to include the tests
862 used for 19 above in the standard set of tests.
863
864 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
865 reference to a subpattern the other side of a comment that contains an
866 opening parenthesis caused either an internal compiling error, or a
867 reference to the wrong subpattern.
868
869
870 Version 8.02 19-Mar-2010
871 ------------------------
872
873 1. The Unicode data tables have been updated to Unicode 5.2.0.
874
875 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
876 configured.
877
878 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
879 original author of that file, following a query about its status.
880
881 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
882 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
883
884 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
885 quantifier applied to a forward-referencing subroutine call, could compile
886 incorrect code or give the error "internal error: previously-checked
887 referenced subpattern not found".
888
889 6. Both MS Visual Studio and Symbian OS have problems with initializing
890 variables to point to external functions. For these systems, therefore,
891 pcre_malloc etc. are now initialized to local functions that call the
892 relevant global functions.
893
894 7. There were two entries missing in the vectors called coptable and poptable
895 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
896 I've fixed the data, and added a kludgy way of testing at compile time that
897 the lengths are correct (equal to the number of opcodes).
898
899 8. Following on from 7, I added a similar kludge to check the length of the
900 eint vector in pcreposix.c.
901
902 9. Error texts for pcre_compile() are held as one long string to avoid too
903 much relocation at load time. To find a text, the string is searched,
904 counting zeros. There was no check for running off the end of the string,
905 which could happen if a new error number was added without updating the
906 string.
907
908 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
909
910 11. \K was not working if it appeared in an atomic group or in a group that
911 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
912 \K is "not well defined" if used in an assertion. PCRE now accepts it if
913 the assertion is positive, but not if it is negative.
914
915 12. Change 11 fortuitously reduced the size of the stack frame used in the
916 "match()" function of pcre_exec.c by one pointer. Forthcoming
917 implementation of support for (*MARK) will need an extra pointer on the
918 stack; I have reserved it now, so that the stack frame size does not
919 decrease.
920
921 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
922 item in branch that calls a recursion is a subroutine call - as in the
923 second branch in the above example - was incorrectly given the compile-
924 time error "recursive call could loop indefinitely" because pcre_compile()
925 was not correctly checking the subroutine for matching a non-empty string.
926
927 14. The checks for overrunning compiling workspace could trigger after an
928 overrun had occurred. This is a "should never occur" error, but it can be
929 triggered by pathological patterns such as hundreds of nested parentheses.
930 The checks now trigger 100 bytes before the end of the workspace.
931
932 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
933
934
935 Version 8.01 19-Jan-2010
936 ------------------------
937
938 1. If a pattern contained a conditional subpattern with only one branch (in
939 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
940 computed the wrong minimum data length (which is of course zero for such
941 subpatterns). This could cause incorrect "no match" results.
942
943 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
944 the pattern is reset in the first branch, pcre_compile() failed with
945 "internal error: code overflow at offset...". This happened only when
946 the reset was to the original external option setting. (An optimization
947 abstracts leading options settings into an external setting, which was the
948 cause of this.)
949
950 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
951 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
952 assertion pattern did not match (meaning that the assertion was true), it
953 was incorrectly treated as false if the SKIP had been reached during the
954 matching. This also applied to assertions used as conditions.
955
956 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
957 assertion subpattern, including such a pattern used as a condition,
958 unpredictable results occurred, instead of the error return
959 PCRE_ERROR_DFA_UITEM.
960
961 5. The C++ GlobalReplace function was not working like Perl for the special
962 situation when an empty string is matched. It now does the fancy magic
963 stuff that is necessary.
964
965 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
966 removed. (These were left over from very, very early versions of PCRE.)
967
968 7. Some cosmetic changes to the code to make life easier when compiling it
969 as part of something else:
970
971 (a) Change DEBUG to PCRE_DEBUG.
972
973 (b) In pcre_compile(), rename the member of the "branch_chain" structure
974 called "current" as "current_branch", to prevent a collision with the
975 Linux macro when compiled as a kernel module.
976
977 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
978 prevent a collision with the Linux macro when compiled as a kernel
979 module.
980
981 8. In pcre_compile() there are some checks for integer overflows that used to
982 cast potentially large values to (double). This has been changed to that
983 when building, a check for int64_t is made, and if it is found, it is used
984 instead, thus avoiding the use of floating point arithmetic. (There is no
985 other use of FP in PCRE.) If int64_t is not found, the fallback is to
986 double.
987
988 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
989 2005 (difference between two addresses compared to an unsigned value).
990
991 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
992 custom one, because of the following reported problem in Windows:
993
994 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
995 under Win32.
996 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
997 therefore missing the function definition.
998 - The compiler thus generates a "C" signature for the test function.
999 - The linker fails to find the "C" function.
1000 - PCRE fails to configure if asked to do so against libbz2.
1001
1002 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1003 messages were output:
1004
1005 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1006 rerunning libtoolize, to keep the correct libtool macros in-tree.
1007 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1008
1009 I have done both of these things.
1010
1011 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1012 most of the time, it *can* run out if it is given a pattern that contains a
1013 runaway infinite recursion. I updated the discussion in the pcrestack man
1014 page.
1015
1016 13. Now that we have gone to the x.xx style of version numbers, the minor
1017 version may start with zero. Using 08 or 09 is a bad idea because users
1018 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1019 interpreted as invalid octal numbers. I've updated the previous comment in
1020 configure.ac, and also added a check that gives an error if 08 or 09 are
1021 used.
1022
1023 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1024 causing partial matching to fail when the end of the subject matched \W
1025 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1026
1027 15. There were some discrepancies between the declarations in pcre_internal.h
1028 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1029 their definitions. The declarations used "const uschar *" and the
1030 definitions used USPTR. Even though USPTR is normally defined as "const
1031 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1032 reported that: "This difference in casting confuses some C++ compilers, for
1033 example, SunCC recognizes above declarations as different functions and
1034 generates broken code for hbpcre." I have changed the declarations to use
1035 USPTR.
1036
1037 16. GNU libtool is named differently on some systems. The autogen.sh script now
1038 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1039 (FreeBSD).
1040
1041 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1042 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1043 comment: "Figure out how to create a longlong from a string: strtoll and
1044 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1045 instance, but it only takes 2 args instead of 3!"
1046
1047 18. A subtle bug concerned with back references has been fixed by a change of
1048 specification, with a corresponding code fix. A pattern such as
1049 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1050 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1051 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1052 same bug. Such groups have to be quantified to be useful, or contained
1053 inside another quantified group. (If there's no repetition, the reference
1054 can never match.) The problem arises because, having left the group and
1055 moved on to the rest of the pattern, a later failure that backtracks into
1056 the group uses the captured value from the final iteration of the group
1057 rather than the correct earlier one. I have fixed this in PCRE by forcing
1058 any group that contains a reference to itself to be an atomic group; that
1059 is, there cannot be any backtracking into it once it has completed. This is
1060 similar to recursive and subroutine calls.
1061
1062
1063 Version 8.00 19-Oct-09
1064 ----------------------
1065
1066 1. The table for translating pcre_compile() error codes into POSIX error codes
1067 was out-of-date, and there was no check on the pcre_compile() error code
1068 being within the table. This could lead to an OK return being given in
1069 error.
1070
1071 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1072 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1073 in a Windows environment.
1074
1075 3. The pcregrep --count option prints the count for each file even when it is
1076 zero, as does GNU grep. However, pcregrep was also printing all files when
1077 --files-with-matches was added. Now, when both options are given, it prints
1078 counts only for those files that have at least one match. (GNU grep just
1079 prints the file name in this circumstance, but including the count seems
1080 more useful - otherwise, why use --count?) Also ensured that the
1081 combination -clh just lists non-zero counts, with no names.
1082
1083 4. The long form of the pcregrep -F option was incorrectly implemented as
1084 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1085 but it seems right to fix it, and I didn't think it was worth preserving
1086 the old behaviour.
1087
1088 5. The command line items --regex=pattern and --regexp=pattern were not
1089 recognized by pcregrep, which required --regex pattern or --regexp pattern
1090 (with a space rather than an '='). The man page documented the '=' forms,
1091 which are compatible with GNU grep; these now work.
1092
1093 6. No libpcreposix.pc file was created for pkg-config; there was just
1094 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1095
1096 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1097 when UCP support is not needed, by modifying the Python script that
1098 generates it from Unicode data files. This should not matter if the module
1099 is correctly used as a library, but I received one complaint about 50K of
1100 unwanted data. My guess is that the person linked everything into his
1101 program rather than using a library. Anyway, it does no harm.
1102
1103 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1104 was a minimum greater than 1 for a wide character in a possessive
1105 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1106 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1107 character. Chaos in the form of incorrect output or a compiling loop could
1108 result.
1109
1110 9. The restrictions on what a pattern can contain when partial matching is
1111 requested for pcre_exec() have been removed. All patterns can now be
1112 partially matched by this function. In addition, if there are at least two
1113 slots in the offset vector, the offset of the earliest inspected character
1114 for the match and the offset of the end of the subject are set in them when
1115 PCRE_ERROR_PARTIAL is returned.
1116
1117 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1118 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1119 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1120 and may be more useful for multi-segment matching.
1121
1122 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1123 used to be given if ever the end of the subject was reached; now it is
1124 given only if matching could not proceed because another character was
1125 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1126 string "Z", which now yields "no match" instead of "partial match". In the
1127 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1128 final character ended with (*FAIL).
1129
1130 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1131 if the pattern had a "must contain" character that was already found in the
1132 earlier partial match, unless partial matching was again requested. For
1133 example, with the pattern /dog.(body)?/, the "must contain" character is
1134 "g". If the first part-match was for the string "dog", restarting with
1135 "sbody" failed. This bug has been fixed.
1136
1137 13. The string returned by pcre_dfa_exec() after a partial match has been
1138 changed so that it starts at the first inspected character rather than the
1139 first character of the match. This makes a difference only if the pattern
1140 starts with a lookbehind assertion or \b or \B (\K is not supported by
1141 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1142 matching functions compatible, and I think it's the right thing to do.
1143
1144 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1145 so that the demonstration program is easily available in environments where
1146 PCRE has not been installed from source.
1147
1148 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1149 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1150 library.
1151
1152 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1153 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1154 is not the first non-POSIX option to be added. Clearly some people find
1155 these options useful.
1156
1157 17. If a caller to the POSIX matching function regexec() passes a non-zero
1158 value for nmatch with a NULL value for pmatch, the value of
1159 nmatch is forced to zero.
1160
1161 18. RunGrepTest did not have a test for the availability of the -u option of
1162 the diff command, as RunTest does. It now checks in the same way as
1163 RunTest, and also checks for the -b option.
1164
1165 19. If an odd number of negated classes containing just a single character
1166 interposed, within parentheses, between a forward reference to a named
1167 subpattern and the definition of the subpattern, compilation crashed with
1168 an internal error, complaining that it could not find the referenced
1169 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1170 [The bug was that it was starting one character too far in when skipping
1171 over the character class, thus treating the ] as data rather than
1172 terminating the class. This meant it could skip too much.]
1173
1174 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1175 /g option in pcretest when the pattern contains \K, which makes it possible
1176 to have an empty string match not at the start, even when the pattern is
1177 anchored. Updated pcretest and pcredemo to use this option.
1178
1179 21. If the maximum number of capturing subpatterns in a recursion was greater
1180 than the maximum at the outer level, the higher number was returned, but
1181 with unset values at the outer level. The correct (outer level) value is
1182 now given.
1183
1184 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1185 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1186 make it do so. The string so far is captured, making this feature
1187 compatible with Perl.
1188
1189 23. The tests have been re-organized, adding tests 11 and 12, to make it
1190 possible to check the Perl 5.10 features against Perl 5.10.
1191
1192 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1193 pattern matches a fixed length string. PCRE did not allow this; now it
1194 does. Neither allows recursion.
1195
1196 25. I finally figured out how to implement a request to provide the minimum
1197 length of subject string that was needed in order to match a given pattern.
1198 (It was back references and recursion that I had previously got hung up
1199 on.) This code has now been added to pcre_study(); it finds a lower bound
1200 to the length of subject needed. It is not necessarily the greatest lower
1201 bound, but using it to avoid searching strings that are too short does give
1202 some useful speed-ups. The value is available to calling programs via
1203 pcre_fullinfo().
1204
1205 26. While implementing 25, I discovered to my embarrassment that pcretest had
1206 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1207 study optimizations had never been tested with that matching function.
1208 Oops. What is worse, even when it was passed study data, there was a bug in
1209 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1210 were also very few tests of studied patterns with pcre_dfa_exec().
1211
1212 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1213 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1214 on the other side of the coin, they are no longer allowed to have different
1215 names, because these cannot be distinguished in PCRE, and this has caused
1216 confusion. (This is a difference from Perl.)
1217
1218 28. When duplicate subpattern names are present (necessarily with different
1219 numbers, as required by 27 above), and a test is made by name in a
1220 conditional pattern, either for a subpattern having been matched, or for
1221 recursion in such a pattern, all the associated numbered subpatterns are
1222 tested, and the overall condition is true if the condition is true for any
1223 one of them. This is the way Perl works, and is also more like the way
1224 testing by number works.
1225
1226
1227 Version 7.9 11-Apr-09
1228 ---------------------
1229
1230 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1231 (pcretest), all targets were linked against these libraries. This included
1232 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1233 libraries. This caused unwanted dependencies to be created. This problem
1234 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1235 pcretest is linked with readline.
1236
1237 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1238 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1239 moved outside it again, because FALSE and TRUE are already defined in AIX,
1240 but BOOL is not.
1241
1242 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1243 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1244
1245 4. The pcregrep documentation said spaces were inserted as well as colons (or
1246 hyphens) following file names and line numbers when outputting matching
1247 lines. This is not true; no spaces are inserted. I have also clarified the
1248 wording for the --colour (or --color) option.
1249
1250 5. In pcregrep, when --colour was used with -o, the list of matching strings
1251 was not coloured; this is different to GNU grep, so I have changed it to be
1252 the same.
1253
1254 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1255 each matching line was coloured. Now it goes on to look for further matches
1256 of any of the test patterns, which is the same behaviour as GNU grep.
1257
1258 7. A pattern that could match an empty string could cause pcregrep to loop; it
1259 doesn't make sense to accept an empty string match in pcregrep, so I have
1260 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1261 seems to be how GNU grep behaves.
1262
1263 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1264 start or after a newline", because the conditional assertion was not being
1265 correctly handled. The rule now is that both the assertion and what follows
1266 in the first alternative must satisfy the test.
1267
1268 9. If auto-callout was enabled in a pattern with a conditional group whose
1269 condition was an assertion, PCRE could crash during matching, both with
1270 pcre_exec() and pcre_dfa_exec().
1271
1272 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1273 used for matching.
1274
1275 11. Unicode property support in character classes was not working for
1276 characters (bytes) greater than 127 when not in UTF-8 mode.
1277
1278 12. Added the -M command line option to pcretest.
1279
1280 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1281
1282 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1283
1284 16. Added comments and documentation about mis-use of no_arg in the C++
1285 wrapper.
1286
1287 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1288 from Martin Jerabek that uses macro names for all relevant character and
1289 string constants.
1290
1291 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1292 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1293 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1294 these, but not everybody uses configure.
1295
1296 19. A conditional group that had only one branch was not being correctly
1297 recognized as an item that could match an empty string. This meant that an
1298 enclosing group might also not be so recognized, causing infinite looping
1299 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1300 with the subject "ab", where knowledge that the repeated group can match
1301 nothing is needed in order to break the loop.
1302
1303 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1304 exec(), but without supplying a callout function, matching went wrong.
1305
1306 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1307 leak if the size of the offset vector was greater than 30. When the vector
1308 is smaller, the saved offsets during recursion go onto a local stack
1309 vector, but for larger vectors malloc() is used. It was failing to free
1310 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1311 error, in fact).
1312
1313 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1314 heapframe that is used only when UTF-8 support is enabled. This caused no
1315 problem, but was untidy.
1316
1317 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1318 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1319 included within another project.
1320
1321 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1322 slightly modified by me:
1323
1324 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1325 not building pcregrep.
1326
1327 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1328 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1329
1330 25. Forward references, both numeric and by name, in patterns that made use of
1331 duplicate group numbers, could behave incorrectly or give incorrect errors,
1332 because when scanning forward to find the reference group, PCRE was not
1333 taking into account the duplicate group numbers. A pattern such as
1334 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1335
1336 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1337 the feature of a custom pointer more persuasive (as requested by a user).
1338
1339 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1340 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1341 pre-defined.
1342
1343 28. Added support for (*UTF8) at the start of a pattern.
1344
1345 29. Arrange for flags added by the "release type" setting in CMake to be shown
1346 in the configuration summary.
1347
1348
1349 Version 7.8 05-Sep-08
1350 ---------------------
1351
1352 1. Replaced UCP searching code with optimized version as implemented for Ad
1353 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1354 stage table and inline lookup instead of a function, giving speed ups of 2
1355 to 5 times on some simple patterns that I tested. Permission was given to
1356 distribute the MultiStage2.py script that generates the tables (it's not in
1357 the tarball, but is in the Subversion repository).
1358
1359 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1360 scripts.
1361
1362 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1363 a group with a zero qualifier. The result of the study could be incorrect,
1364 or the function might crash, depending on the pattern.
1365
1366 4. Caseless matching was not working for non-ASCII characters in back
1367 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1368 It now works when Unicode Property Support is available.
1369
1370 5. In pcretest, an escape such as \x{de} in the data was always generating
1371 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1372 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1373 truncation.
1374
1375 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1376
1377 7. Added two (int) casts to pcregrep when printing the difference of two
1378 pointers, in case they are 64-bit values.
1379
1380 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1381 test 2 if it fails.
1382
1383 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1384 and a #define of that name to empty if it is not externally set. This is to
1385 allow users of MSVC to set it if necessary.
1386
1387 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1388 the convenience functions in the pcre_get.c source file.
1389
1390 11. An option change at the start of a pattern that had top-level alternatives
1391 could cause overwriting and/or a crash. This command provoked a crash in
1392 some environments:
1393
1394 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1395
1396 This potential security problem was recorded as CVE-2008-2371.
1397
1398 12. For a pattern where the match had to start at the beginning or immediately
1399 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1400 pcre_dfa_exec() could read past the end of the passed subject if there was
1401 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1402 pcretest so that it places the subject at the end of its malloc-ed buffer.
1403
1404 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1405 exec() might read past the end of the data buffer in UTF-8 mode.
1406
1407 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1408 the data contained the byte 0x85 as part of a UTF-8 character within its
1409 first line. This applied both to normal and DFA matching.
1410
1411 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1412 /^[^d]*?$/8 failed to match "abc".
1413
1414 16. Added a missing copyright notice to pcrecpp_internal.h.
1415
1416 17. Make it more clear in the documentation that values returned from
1417 pcre_exec() in ovector are byte offsets, not character counts.
1418
1419 18. Tidied a few places to stop certain compilers from issuing warnings.
1420
1421 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1422 supplied by Stefan Weber. I made a further small update for 7.8 because
1423 there is a change of source arrangements: the pcre_searchfuncs.c module is
1424 replaced by pcre_ucd.c.
1425
1426
1427 Version 7.7 07-May-08
1428 ---------------------
1429
1430 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1431 a string to a long long, pretend we don't even have a long long." This is
1432 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1433
1434 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1435 pre-7.6 versions, which defined a global no_arg variable instead of putting
1436 it in the RE class. (See also #8 below.)
1437
1438 3. Remove a line of dead code, identified by coverity and reported by Nuno
1439 Lopes.
1440
1441 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1442
1443 (1) The include/exclude patterns were being applied to the whole pathnames
1444 of files, instead of just to the final components.
1445
1446 (2) If there was more than one level of directory, the subdirectories were
1447 skipped unless they satisfied the include/exclude conditions. This is
1448 inconsistent with GNU grep (and could even be seen as contrary to the
1449 pcregrep specification - which I improved to make it absolutely clear).
1450 The action now is always to scan all levels of directory, and just
1451 apply the include/exclude patterns to regular files.
1452
1453 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1454 --exclude_dir in the tests to avoid scanning .svn directories.
1455
1456 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1457 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1458 doesn't support NULs in patterns.
1459
1460 7. Added some missing "const"s to declarations of static tables in
1461 pcre_compile.c and pcre_dfa_exec.c.
1462
1463 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1464 caused by fix #2 above. (Subsequently also a second patch to fix the
1465 first patch. And a third patch - this was a messy problem.)
1466
1467 9. Applied Craig's patch to remove the use of push_back().
1468
1469 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1470 matching function regexec().
1471
1472 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1473 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1474 references. PCRE supports relative numbers with this syntax (I don't think
1475 Oniguruma does).
1476
1477 12. Previously, a group with a zero repeat such as (...){0} was completely
1478 omitted from the compiled regex. However, this means that if the group
1479 was called as a subroutine from elsewhere in the pattern, things went wrong
1480 (an internal error was given). Such groups are now left in the compiled
1481 pattern, with a new opcode that causes them to be skipped at execution
1482 time.
1483
1484 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1485 to the way PCRE behaves:
1486
1487 (a) A lone ] character is dis-allowed (Perl treats it as data).
1488
1489 (b) A back reference to an unmatched subpattern matches an empty string
1490 (Perl fails the current match path).
1491
1492 (c) A data ] in a character class must be notated as \] because if the
1493 first data character in a class is ], it defines an empty class. (In
1494 Perl it is not possible to have an empty class.) The empty class []
1495 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1496 The negative empty class [^] matches any one character, independently
1497 of the DOTALL setting.
1498
1499 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1500 non-existent subpattern following a character class starting with ']' and
1501 containing () gave an internal compiling error instead of "reference to
1502 non-existent subpattern". Fortunately, when the pattern did exist, the
1503 compiled code was correct. (When scanning forwards to check for the
1504 existencd of the subpattern, it was treating the data ']' as terminating
1505 the class, so got the count wrong. When actually compiling, the reference
1506 was subsequently set up correctly.)
1507
1508 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1509 it was being rejected as not supported by pcre_dfa_exec(), even though
1510 other assertions are supported. I have made pcre_dfa_exec() support
1511 (*FAIL).
1512
1513 16. The implementation of 13c above involved the invention of a new opcode,
1514 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1515 cannot be changed at match time, I realized I could make a small
1516 improvement to matching performance by compiling OP_ALLANY instead of
1517 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1518 on the OP_ANY path.
1519
1520 17. Compiling pcretest on Windows with readline support failed without the
1521 following two fixes: (1) Make the unistd.h include conditional on
1522 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1523
1524 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1525 ncurses library to be included for pcretest when ReadLine support is
1526 requested, but also to allow for it to be overridden. This patch came from
1527 Daniel Bergström.
1528
1529 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1530 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1531 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1532 spotting this.
1533
1534
1535 Version 7.6 28-Jan-08
1536 ---------------------
1537
1538 1. A character class containing a very large number of characters with
1539 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1540 overflow.
1541
1542 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1543 HAVE_LONG_LONG is not defined.
1544
1545 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1546 bring them up to date and include new features. This patch includes:
1547
1548 - Fixed PH's badly added libz and libbz2 support.
1549 - Fixed a problem with static linking.
1550 - Added pcredemo. [But later removed - see 7 below.]
1551 - Fixed dftables problem and added an option.
1552 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1553 HAVE_LONG_LONG.
1554 - Added readline support for pcretest.
1555 - Added an listing of the option settings after cmake has run.
1556
1557 4. A user submitted a patch to Makefile that makes it easy to create
1558 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1559 Makefile.am that cause it to include this special target, without
1560 affecting anything else. Note that the same mingw target plus all
1561 the other distribution libraries and programs are now supported
1562 when configuring with CMake (see 6 below) instead of with
1563 Configure/Make.
1564
1565 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1566 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1567 exported in the Windows port". It has not yet been confirmed that the patch
1568 solves the problem, but it does no harm.
1569
1570 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1571 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1572 with CMake, and also correct the comment about stack recursion.
1573
1574 7. Remove the automatic building of pcredemo from the ./configure system and
1575 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1576 of a program that users should build themselves after PCRE is installed, so
1577 building it automatically is not really right. What is more, it gave
1578 trouble in some build environments.
1579
1580 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1581
1582
1583 Version 7.5 10-Jan-08
1584 ---------------------
1585
1586 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1587 values in parens when parsing an RE using the C++ wrapper."
1588
1589 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1590 Characters greater than 255 were excluded from the class instead of being
1591 included.
1592
1593 3. The same bug as (2) above applied to negated POSIX classes such as
1594 [:^space:].
1595
1596 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1597 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1598 I have changed it.
1599
1600 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1601 first named subpattern) and a construct such as (?&a) would reference the
1602 first named subpattern whose name started with "a" (in other words, the
1603 length check was missing). Both these problems are fixed. "Subpattern name
1604 expected" is now given for (?&) (a zero-length name), and this patch also
1605 makes it give the same error for \k'' (previously it complained that that
1606 was a reference to a non-existent subpattern).
1607
1608 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1609 this is right because (?- can be followed by option settings as well as by
1610 digits. I have, however, made the messages clearer.
1611
1612 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1613 than the number used in the conditional) now cause a compile-time error.
1614 This is actually not compatible with Perl, which accepts such patterns, but
1615 treats the conditional as always being FALSE (as PCRE used to), but it
1616 seems to me that giving a diagnostic is better.
1617
1618 8. Change "alphameric" to the more common word "alphanumeric" in comments
1619 and messages.
1620
1621 9. Fix two occurrences of "backslash" in comments that should have been
1622 "backspace".
1623
1624 10. Remove two redundant lines of code that can never be obeyed (their function
1625 was moved elsewhere).
1626
1627 11. The program that makes PCRE's Unicode character property table had a bug
1628 which caused it to generate incorrect table entries for sequences of
1629 characters that have the same character type, but are in different scripts.
1630 It amalgamated them into a single range, with the script of the first of
1631 them. In other words, some characters were in the wrong script. There were
1632 thirteen such cases, affecting characters in the following ranges:
1633
1634 U+002b0 - U+002c1
1635 U+0060c - U+0060d
1636 U+0061e - U+00612
1637 U+0064b - U+0065e
1638 U+0074d - U+0076d
1639 U+01800 - U+01805
1640 U+01d00 - U+01d77
1641 U+01d9b - U+01dbf
1642 U+0200b - U+0200f
1643 U+030fc - U+030fe
1644 U+03260 - U+0327f
1645 U+0fb46 - U+0fbb1
1646 U+10450 - U+1049d
1647
1648 12. The -o option (show only the matching part of a line) for pcregrep was not
1649 compatible with GNU grep in that, if there was more than one match in a
1650 line, it showed only the first of them. It now behaves in the same way as
1651 GNU grep.
1652
1653 13. If the -o and -v options were combined for pcregrep, it printed a blank
1654 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1655 does the same. The return code can be used to tell if there were any
1656 non-matching lines.
1657
1658 14. Added --file-offsets and --line-offsets to pcregrep.
1659
1660 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1661 infinitely looping recursion. The bug was that positive lookaheads were not
1662 being skipped when checking for a possible empty match (negative lookaheads
1663 and both kinds of lookbehind were skipped).
1664
1665 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1666 inclusion of <windows.h> to before rather than after the definition of
1667 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1668
1669 17. Specifying a possessive quantifier with a specific limit for a Unicode
1670 character property caused pcre_compile() to compile bad code, which led at
1671 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1672 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1673 caused the error; without that there was no problem.
1674
1675 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1676
1677 19. Added --enable-pcretest-libreadline.
1678
1679 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1680 RE::GlobalReplace(). As a result, the number of replacements returned was
1681 double what it should be. I removed one of the increments, but Craig sent a
1682 later patch that removed the other one (the right fix) and added unit tests
1683 that check the return values (which was not done before).
1684
1685 21. Several CMake things:
1686
1687 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1688 the names libpcre and libpcreposix, not just pcre and pcreposix.
1689
1690 (2) The above change means that pcretest and pcregrep are now correctly
1691 linked with the newly-built libraries, not previously installed ones.
1692
1693 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1694
1695 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1696 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1697 UTF-8 newline character). The key issue is that the pattern starts .*;
1698 this means that the match must be either at the beginning, or after a
1699 newline. The bug was in the code for advancing after a failed match and
1700 checking that the new position followed a newline. It was not taking
1701 account of UTF-8 characters correctly.
1702
1703 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1704 character classes. PCRE was not treating the sequence [:...:] as a
1705 character class unless the ... were all letters. Perl, however, seems to
1706 allow any characters between [: and :], though of course it rejects as
1707 unknown any "names" that contain non-letters, because all the known class
1708 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1709 for example, whereas PCRE did not - it did not recognize a POSIX character
1710 class. This seemed a bit dangerous, so the code has been changed to be
1711 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1712 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1713 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1714 Perl does, and where it didn't before.
1715
1716 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1717 Windows environments %n is disabled by default.
1718
1719
1720 Version 7.4 21-Sep-07
1721 ---------------------
1722
1723 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1724 means that a class such as [\s] counted as "explicit reference to CR or
1725 LF". That isn't really right - the whole point of the change was to try to
1726 help when there was an actual mention of one of the two characters. So now
1727 the change happens only if \r or \n (or a literal CR or LF) character is
1728 encountered.
1729
1730 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1731 of both had grown to the point where there were only 3 bits left.
1732 Fortunately, there was spare space in the data structure, and so I have
1733 moved the internal flags into a new 16-bit field to free up more option
1734 bits.
1735
1736 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1737 but did not set the internal JCHANGED flag - either of these is enough to
1738 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1739 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1740 start sets both bits.
1741
1742 4. Added options (at build time, compile time, exec time) to change \R from
1743 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1744
1745 5. doc/pcresyntax.html was missing from the distribution.
1746
1747 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1748 compatibility, even though it is no longer used.
1749
1750 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1751 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1752 windows.h file is present (where different names are used). [This was
1753 reversed later after testing - see 16 below.]
1754
1755 8. Changed all #include <config.h> to #include "config.h". There were also
1756 some further <pcre.h> cases that I changed to "pcre.h".
1757
1758 9. When pcregrep was used with the --colour option, it missed the line ending
1759 sequence off the lines that it output.
1760
1761 10. It was pointed out to me that arrays of string pointers cause lots of
1762 relocations when a shared library is dynamically loaded. A technique of
1763 using a single long string with a table of offsets can drastically reduce
1764 these. I have refactored PCRE in four places to do this. The result is
1765 dramatic:
1766
1767 Originally: 290
1768 After changing UCP table: 187
1769 After changing error message table: 43
1770 After changing table of "verbs" 36
1771 After changing table of Posix names 22
1772
1773 Thanks to the folks working on Gregex for glib for this insight.
1774
1775 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1776 unicode-properties was also set.
1777
1778 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1779
1780 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1781 checked only for CRLF.
1782
1783 14. Added casts to pcretest.c to avoid compiler warnings.
1784
1785 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1786
1787 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1788 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1789 entirely. This removes changes made in 7 above.
1790
1791 17. The CMake files have been updated, and there is now more information about
1792 building with CMake in the NON-UNIX-USE document.
1793
1794
1795 Version 7.3 28-Aug-07
1796 ---------------------
1797
1798 1. In the rejigging of the build system that eventually resulted in 7.1, the
1799 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1800 brackets there is not right, since it causes compilers to look for an
1801 installed pcre.h, not the version that is in the source that is being
1802 compiled (which of course may be different). I have changed it back to:
1803
1804 #include "pcre.h"
1805
1806 I have a vague recollection that the change was concerned with compiling in
1807 different directories, but in the new build system, that is taken care of
1808 by the VPATH setting the Makefile.
1809
1810 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1811 when the subject happened to end in the byte 0x85 (e.g. if the last
1812 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1813 characters but of course it shouldn't be taken as a newline when it is part
1814 of another character. The bug was that, for an unlimited repeat of . in
1815 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1816 characters when looking for a newline.
1817
1818 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1819
1820 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1821 in debug output.
1822
1823 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1824 long printing in the pcrecpp unittest when running under MinGW.
1825
1826 6. ESC_K was left out of the EBCDIC table.
1827
1828 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1829 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1830 limit also applies to "virtual nesting" when a pattern is recursive, and in
1831 this case 1000 isn't so big. I have been able to remove this limit at the
1832 expense of backing off one optimization in certain circumstances. Normally,
1833 when pcre_exec() would call its internal match() function recursively and
1834 immediately return the result unconditionally, it uses a "tail recursion"
1835 feature to save stack. However, when a subpattern that can match an empty
1836 string has an unlimited repetition quantifier, it no longer makes this
1837 optimization. That gives it a stack frame in which to save the data for
1838 checking that an empty string has been matched. Previously this was taken
1839 from the 1000-entry workspace that had been reserved. So now there is no
1840 explicit limit, but more stack is used.
1841
1842 8. Applied Daniel's patches to solve problems with the import/export magic
1843 syntax that is required for Windows, and which was going wrong for the
1844 pcreposix and pcrecpp parts of the library. These were overlooked when this
1845 problem was solved for the main library.
1846
1847 9. There were some crude static tests to avoid integer overflow when computing
1848 the size of patterns that contain repeated groups with explicit upper
1849 limits. As the maximum quantifier is 65535, the maximum group length was
1850 set at 30,000 so that the product of these two numbers did not overflow a
1851 32-bit integer. However, it turns out that people want to use groups that
1852 are longer than 30,000 bytes (though not repeat them that many times).
1853 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1854 made it possible to implement the integer overflow checks in a much more
1855 dynamic way, which I have now done. The artificial limitation on group
1856 length has been removed - we now have only the limit on the total length of
1857 the compiled pattern, which depends on the LINK_SIZE setting.
1858
1859 10. Fixed a bug in the documentation for get/copy named substring when
1860 duplicate names are permitted. If none of the named substrings are set, the
1861 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1862 empty string.
1863
1864 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1865 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1866 because the ] is interpreted as the first data character and the
1867 terminating ] is not found. PCRE has been made compatible with Perl in this
1868 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1869 cause memory overwriting.
1870
1871 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1872 string has been matched (to stop an infinite loop). It was not recognizing
1873 a conditional subpattern that could match an empty string if that
1874 subpattern was within another subpattern. For example, it looped when
1875 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1876 condition was not nested. This bug has been fixed.
1877
1878 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1879 past the start of the subject in the presence of bytes with the top bit
1880 set, for example "\x8aBCD".
1881
1882 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1883 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1884
1885 14. Optimized (?!) to (*FAIL).
1886
1887 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1888 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1889 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1890 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1891 does: it's just the validity check that is more restrictive.
1892
1893 16. Inserted checks for integer overflows during escape sequence (backslash)
1894 processing, and also fixed erroneous offset values for syntax errors during
1895 backslash processing.
1896
1897 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1898 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1899
1900 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1901 caused an overrun.
1902
1903 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1904 something other than just ASCII characters) inside a group that had an
1905 unlimited repeat caused a loop at compile time (while checking to see
1906 whether the group could match an empty string).
1907
1908 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1909 [\P{Any}] did so. (Error in the code for printing property names.)
1910
1911 21. An orphan \E inside a character class could cause a crash.
1912
1913 22. A repeated capturing bracket such as (A)? could cause a wild memory
1914 reference during compilation.
1915
1916 23. There are several functions in pcre_compile() that scan along a compiled
1917 expression for various reasons (e.g. to see if it's fixed length for look
1918 behind). There were bugs in these functions when a repeated \p or \P was
1919 present in the pattern. These operators have additional parameters compared
1920 with \d, etc, and these were not being taken into account when moving along
1921 the compiled data. Specifically:
1922
1923 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1924 length.
1925
1926 (b) An item such as \pL+ within a repeated group could cause crashes or
1927 loops.
1928
1929 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1930 "reference to non-existent subpattern" error.
1931
1932 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1933
1934 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1935 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1936
1937 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1938
1939 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1940 character were causing crashes (broken optimization).
1941
1942 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1943 \p or \P) caused a compile-time loop.
1944
1945 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1946 break. For example, the unstudied pattern [\r\n]A does not match the string
1947 "\r\nA" because change 7.0/46 below moves the current point on by two
1948 characters after failing to match at the start. However, the pattern \nA
1949 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1950 the same is true. There doesn't seem any very clean way out of this, but
1951 what I have chosen to do makes the common cases work: PCRE now takes note
1952 of whether there can be an explicit match for \r or \n anywhere in the
1953 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1954 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1955 pattern has explicit CR or LF references.
1956
1957 29. Added (*CR) etc for changing newline setting at start of pattern.
1958
1959
1960 Version 7.2 19-Jun-07
1961 ---------------------
1962
1963 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1964 which is apparently normally available under Windows.
1965
1966 2. Re-jig the pcregrep tests with different newline settings in an attempt
1967 to make them independent of the local environment's newline setting.
1968
1969 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1970
1971 4. Some of the "internals" tests were previously cut out when the link size
1972 was not 2, because the output contained actual offsets. The recent new
1973 "Z" feature of pcretest means that these can be cut out, making the tests
1974 usable with all link sizes.
1975
1976 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1977 stack recursion. This gives a massive performance boost under BSD, but just
1978 a small improvement under Linux. However, it saves one field in the frame
1979 in all cases.
1980
1981 6. Added more features from the forthcoming Perl 5.10:
1982
1983 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1984 recursion call. It refers to the nth most recently opened parentheses.
1985
1986 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1987 to be opened parentheses.
1988
1989 (c) Conditions that refer to capturing parentheses can be specified
1990 relatively, for example, (?(-2)... or (?(+3)...
1991
1992 (d) \K resets the start of the current match so that everything before
1993 is not part of it.
1994
1995 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1996
1997 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1998 reference syntax.
1999
2000 (g) (?| introduces a group in which the numbering of parentheses in each
2001 alternative starts with the same number.
2002
2003 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2004
2005 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2006 PCRE_INFO_JCHANGED.
2007
2008 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2009 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2010 for detecting groups that can match an empty string.
2011
2012 9. A pattern with a very large number of alternatives (more than several
2013 hundred) was running out of internal workspace during the pre-compile
2014 phase, where pcre_compile() figures out how much memory will be needed. A
2015 bit of new cunning has reduced the workspace needed for groups with
2016 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2017 workspace instead of running out of the 4096 that are available.
2018
2019 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2020
2021 11. Applied patch from Google to remove an optimization that didn't quite work.
2022 The report of the bug said:
2023
2024 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2025 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2026 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2027
2028 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2029 it matched the wrong number of bytes.
2030
2031
2032 Version 7.1 24-Apr-07
2033 ---------------------
2034
2035 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2036 that is more "standard", making use of automake and other Autotools. There
2037 is some re-arrangement of the files and adjustment of comments consequent
2038 on this.
2039
2040 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2041 for recursive directory scanning broke on some systems because the files
2042 are not scanned in any specific order and on different systems the order
2043 was different. A call to "sort" has been inserted into RunGrepTest for the
2044 approprate test as a short-term fix. In the longer term there may be an
2045 alternative.
2046
2047 3. I had an email from Eric Raymond about problems translating some of PCRE's
2048 man pages to HTML (despite the fact that I distribute HTML pages, some
2049 people do their own conversions for various reasons). The problems
2050 concerned the use of low-level troff macros .br and .in. I have therefore
2051 removed all such uses from the man pages (some were redundant, some could
2052 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2053 HTML has been updated to handle .nf/.fi and to complain if it encounters
2054 .br or .in.
2055
2056 4. Updated comments in configure.ac that get placed in config.h.in and also
2057 arranged for config.h to be included in the distribution, with the name
2058 config.h.generic, for the benefit of those who have to compile without
2059 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2060
2061 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2062 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2063 makevp.bat for the current PCRE, using the additional files
2064 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2065
2066 6. A Windows user reported a minor discrepancy with test 2, which turned out
2067 to be caused by a trailing space on an input line that had got lost in his
2068 copy. The trailing space was an accident, so I've just removed it.
2069
2070 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2071 that is needed.
2072
2073 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2074 as "const" (a) because they are and (b) because it helps the PHP
2075 maintainers who have recently made a script to detect big data structures
2076 in the php code that should be moved to the .rodata section. I remembered
2077 to update Builducptable as well, so it won't revert if ucptable.h is ever
2078 re-created.
2079
2080 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2081 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2082 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2083 support is not required. This saves 1.5-2K of code, which is important in
2084 some applications.
2085
2086 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2087 so as not to refer to the tables, even though these functions will never be
2088 called when UTF-8 support is disabled. Otherwise there are problems with a
2089 shared library.
2090
2091 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2092
2093 (a) It was defining its arguments as char * instead of void *.
2094
2095 (b) It was assuming that all moves were upwards in memory; this was true
2096 a long time ago when I wrote it, but is no longer the case.
2097
2098 The emulated memove() is provided for those environments that have neither
2099 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2100 is clearly not the case, as these two bugs were recently reported.
2101
2102 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2103 and Detrail to create the HTML documentation, the .txt form of the man
2104 pages, and it removes trailing spaces from listed files. It also creates
2105 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2106 case, it wraps all the #defines with #ifndefs. This script should be run
2107 before "make dist".
2108
2109 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2110 with Unicode property support.
2111
2112 (a) For a maximizing quantifier, if the two different cases of the
2113 character were of different lengths in their UTF-8 codings (there are
2114 some cases like this - I found 11), and the matching function had to
2115 back up over a mixture of the two cases, it incorrectly assumed they
2116 were both the same length.
2117
2118 (b) When PCRE was configured to use the heap rather than the stack for
2119 recursion during matching, it was not correctly preserving the data for
2120 the other case of a UTF-8 character when checking ahead for a match
2121 while processing a minimizing repeat. If the check also involved
2122 matching a wide character, but failed, corruption could cause an
2123 erroneous result when trying to check for a repeat of the original
2124 character.
2125
2126 13. Some tidying changes to the testing mechanism:
2127
2128 (a) The RunTest script now detects the internal link size and whether there
2129 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2130 values substituted by "configure". (The RunGrepTest script already did
2131 this for UTF-8.) The configure.ac script no longer substitutes the
2132 relevant variables.
2133
2134 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2135 with length and offset values. This means that the output is different
2136 for different internal link sizes. Test 2 is skipped for link sizes
2137 other than 2 because of this, bypassing the problem. Unfortunately,
2138 there was also a test in test 3 (the locale tests) that used /B and
2139 failed for link sizes other than 2. Rather than cut the whole test out,
2140 I have added a new /Z option to pcretest that replaces the length and
2141 offset values with spaces. This is now used to make test 3 independent
2142 of link size. (Test 2 will be tidied up later.)
2143
2144 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2145 segmentation fault instead of returning the appropriate error message.
2146
2147 15. In multiline mode when the newline sequence was set to "any", the pattern
2148 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2149 This doesn't seem right; it now treats the CRLF combination as the line
2150 ending, and so does not match in that case. It's only a pattern such as ^$
2151 that would hit this one: something like ^ABC$ would have failed after \r
2152 and then tried again after \r\n.
2153
2154 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2155 in an attempt to make files that differ only in their line terminators
2156 compare equal. This works on Linux.
2157
2158 17. Under certain error circumstances pcregrep might try to free random memory
2159 as it exited. This is now fixed, thanks to valgrind.
2160
2161 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2162 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2163 was because its rules for how to advance for /g after matching an empty
2164 string at the end of a line did not allow for this case. They now check for
2165 it specially.
2166
2167 20. pcretest is supposed to handle patterns and data of any length, by
2168 extending its buffers when necessary. It was getting this wrong when the
2169 buffer for a data line had to be extended.
2170
2171 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2172 CRLF as a newline sequence.
2173
2174 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2175 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2176 I have nevertheless tidied it up.
2177
2178 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2179
2180 24. Added a man page for pcre-config.
2181
2182
2183 Version 7.0 19-Dec-06
2184 ---------------------
2185
2186 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2187 moving to gcc 4.1.1.
2188
2189 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2190 sys/time.h, which is documented as needed for this function. It doesn't
2191 seem to matter on Linux, but it showed up on some releases of OS X.
2192
2193 3. It seems that there are systems where bytes whose values are greater than
2194 127 match isprint() in the "C" locale. The "C" locale should be the
2195 default when a C program starts up. In most systems, only ASCII printing
2196 characters match isprint(). This difference caused the output from pcretest
2197 to vary, making some of the tests fail. I have changed pcretest so that:
2198
2199 (a) When it is outputting text in the compiled version of a pattern, bytes
2200 other than 32-126 are always shown as hex escapes.
2201
2202 (b) When it is outputting text that is a matched part of a subject string,
2203 it does the same, unless a different locale has been set for the match
2204 (using the /L modifier). In this case, it uses isprint() to decide.
2205
2206 4. Fixed a major bug that caused incorrect computation of the amount of memory
2207 required for a compiled pattern when options that changed within the
2208 pattern affected the logic of the preliminary scan that determines the
2209 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2210 that the computed length was too small. The symptoms of this bug were
2211 either the PCRE error "internal error: code overflow" from pcre_compile(),
2212 or a glibc crash with a message such as "pcretest: free(): invalid next
2213 size (fast)". Examples of patterns that provoked this bug (shown in
2214 pcretest format) are:
2215
2216 /(?-x: )/x
2217 /(?x)(?-x: \s*#\s*)/
2218 /((?i)[\x{c0}])/8
2219 /(?i:[\x{c0}])/8
2220
2221 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2222 is now done differently.
2223
2224 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2225 wrapper classes; (b) implement a new function in the C++ scanner that is
2226 more efficient than the old way of doing things because it avoids levels of
2227 recursion in the regex matching; (c) add a paragraph to the documentation
2228 for the FullMatch() function.
2229
2230 6. The escape sequence \n was being treated as whatever was defined as
2231 "newline". Not only was this contrary to the documentation, which states
2232 that \n is character 10 (hex 0A), but it also went horribly wrong when
2233 "newline" was defined as CRLF. This has been fixed.
2234
2235 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2236 was being set to -1 for the "end of line" case (supposedly a value that no
2237 character can have). Though this value is never used (the check for end of
2238 line is "zero bytes in current character"), it caused compiler complaints.
2239 I've changed it to 0xffffffff.
2240
2241 8. In pcre_version.c, the version string was being built by a sequence of
2242 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2243 string (as it is for production releases) called a macro with an empty
2244 argument. The C standard says the result of this is undefined. The gcc
2245 compiler treats it as an empty string (which was what was wanted) but it is
2246 reported that Visual C gives an error. The source has been hacked around to
2247 avoid this problem.
2248
2249 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2250 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2251 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2252 of them did).
2253
2254 10. Originally, pcretest opened its input and output without "b"; then I was
2255 told that "b" was needed in some environments, so it was added for release
2256 5.0 to both the input and output. (It makes no difference on Unix-like
2257 systems.) Later I was told that it is wrong for the input on Windows. I've
2258 now abstracted the modes into two macros, to make it easier to fiddle with
2259 them, and removed "b" from the input mode under Windows.
2260
2261 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2262
2263 12. Added -help and --help to pcretest as an official way of being reminded
2264 of the options.
2265
2266 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2267 and pcrecpp.cc because they annoy compilers at high warning levels.
2268
2269 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2270
2271 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2272 scripts are not C programs :-) and which was not noticed because it works
2273 on Linux.
2274
2275 16. pcretest is supposed to handle any length of pattern and data line (as one
2276 line or as a continued sequence of lines) by extending its input buffer if
2277 necessary. This feature was broken for very long pattern lines, leading to
2278 a string of junk being passed to pcre_compile() if the pattern was longer
2279 than about 50K.
2280
2281 17. I have done a major re-factoring of the way pcre_compile() computes the
2282 amount of memory needed for a compiled pattern. Previously, there was code
2283 that made a preliminary scan of the pattern in order to do this. That was
2284 OK when PCRE was new, but as the facilities have expanded, it has become
2285 harder and harder to keep it in step with the real compile phase, and there
2286 have been a number of bugs (see for example, 4 above). I have now found a
2287 cunning way of running the real compile function in a "fake" mode that
2288 enables it to compute how much memory it would need, while actually only
2289 ever using a few hundred bytes of working memory and without too many
2290 tests of the mode. This should make future maintenance and development
2291 easier. A side effect of this work is that the limit of 200 on the nesting
2292 depth of parentheses has been removed (though this was never a serious
2293 limitation, I suspect). However, there is a downside: pcre_compile() now
2294 runs more slowly than before (30% or more, depending on the pattern). I
2295 hope this isn't a big issue. There is no effect on runtime performance.
2296
2297 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2298 newline (only possible for the last line of a file) and it was a
2299 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2300
2301 19. Added additional timing features to pcretest. (1) The -tm option now times
2302 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2303 separate command line item, by a number that specifies the number of
2304 repeats to use when timing. The default is 50000; this gives better
2305 precision, but takes uncomfortably long for very large patterns.
2306
2307 20. Extended pcre_study() to be more clever in cases where a branch of a
2308 subpattern has no definite first character. For example, (a*|b*)[cd] would
2309 previously give no result from pcre_study(). Now it recognizes that the
2310 first character must be a, b, c, or d.
2311
2312 21. There was an incorrect error "recursive call could loop indefinitely" if
2313 a subpattern (or the entire pattern) that was being tested for matching an
2314 empty string contained only one non-empty item after a nested subpattern.
2315 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2316 incorrectly, because the \d was being skipped in the check.
2317
2318 22. The pcretest program now has a new pattern option /B and a command line
2319 option -b, which is equivalent to adding /B to every pattern. This causes
2320 it to show the compiled bytecode, without the additional information that
2321 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2322 is the same as /B/I).
2323
2324 23. A new optimization is now able automatically to treat some sequences such
2325 as a*b as a*+b. More specifically, if something simple (such as a character
2326 or a simple class like \d) has an unlimited quantifier, and is followed by
2327 something that cannot possibly match the quantified thing, the quantifier
2328 is automatically "possessified".
2329
2330 24. A recursive reference to a subpattern whose number was greater than 39
2331 went wrong under certain circumstances in UTF-8 mode. This bug could also
2332 have affected the operation of pcre_study().
2333
2334 25. Realized that a little bit of performance could be had by replacing
2335 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2336
2337 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2338
2339 27. Possessive quantifiers such as a++ were previously implemented by turning
2340 them into atomic groups such as ($>a+). Now they have their own opcodes,
2341 which improves performance. This includes the automatically created ones
2342 from 23 above.
2343
2344 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2345 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2346 the first matched character to be a colon. This applied both to named and
2347 numbered groups.
2348
2349 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2350
2351 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2352 building PCRE on Windows easier, so I have included it in the distribution.
2353
2354 31. There is now a check in pcretest against a ridiculously large number being
2355 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2356 loop, the loop is abandoned.
2357
2358 32. Forward references to subpatterns in conditions such as (?(2)...) where
2359 subpattern 2 is defined later cause pcre_compile() to search forwards in
2360 the pattern for the relevant set of parentheses. This search went wrong
2361 when there were unescaped parentheses in a character class, parentheses
2362 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2363
2364 33. "Subroutine" calls and backreferences were previously restricted to
2365 referencing subpatterns earlier in the regex. This restriction has now
2366 been removed.
2367
2368 34. Added a number of extra features that are going to be in Perl 5.10. On the
2369 whole, these are just syntactic alternatives for features that PCRE had
2370 previously implemented using the Python syntax or my own invention. The
2371 other formats are all retained for compatibility.
2372
2373 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2374 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2375 also .NET compatible.
2376
2377 (b) A recursion or subroutine call to a named group can now be defined as
2378 (?&name) as well as (?P>name).
2379
2380 (c) A backreference to a named group can now be defined as \k<name> or
2381 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2382 5.10, are also .NET compatible.
2383
2384 (d) A conditional reference to a named group can now use the syntax
2385 (?(<name>) or (?('name') as well as (?(name).
2386
2387 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2388 groups (named and numbered) that are never evaluated inline, but can be
2389 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2390 is always false. There may be only one alternative in such a group.
2391
2392 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2393 as the simple (?(R). The condition is true only if the most recent
2394 recursion is that of the given number or name. It does not search out
2395 through the entire recursion stack.
2396
2397 (g) The escape \gN or \g{N} has been added, where N is a positive or
2398 negative number, specifying an absolute or relative reference.
2399
2400 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2401 some "unreachable code" warnings.
2402
2403 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2404 things, this adds five new scripts.
2405
2406 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2407 There were also incompatibilities regarding the handling of \Q..\E inside
2408 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2409 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2410
2411 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2412 matches an empty string, and forcibly breaks the loop. There were bugs in
2413 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2414 against aaaa the result was just "a" rather than "aaaa", for example. Two
2415 separate and independent bugs (that affected different cases) have been
2416 fixed.
2417
2418 39. Refactored the code to abolish the use of different opcodes for small
2419 capturing bracket numbers. This is a tidy that I avoided doing when I
2420 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2421 The new approach is not only tidier, it makes it possible to reduce the
2422 memory needed to fix the previous bug (38).
2423
2424 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2425 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2426 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2427 mode.
2428
2429 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2430 report.
2431
2432 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2433 copy construction and assignment in the C++ wrapper.
2434
2435 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2436 couple of bugs that could have given wrong results in the "--newline=crlf"
2437 case.
2438
2439 44. Added a number of casts and did some reorganization of signed/unsigned int
2440 variables following suggestions from Dair Grant. Also renamed the variable
2441 "this" as "item" because it is a C++ keyword.
2442
2443 45. Arranged for dftables to add
2444
2445 #include "pcre_internal.h"
2446
2447 to pcre_chartables.c because without it, gcc 4.x may remove the array
2448 definition from the final binary if PCRE is built into a static library and
2449 dead code stripping is activated.
2450
2451 46. For an unanchored pattern, if a match attempt fails at the start of a
2452 newline sequence, and the newline setting is CRLF or ANY, and the next two
2453 characters are CRLF, advance by two characters instead of one.
2454
2455
2456 Version 6.7 04-Jul-06
2457 ---------------------
2458
2459 1. In order to handle tests when input lines are enormously long, pcretest has
2460 been re-factored so that it automatically extends its buffers when
2461 necessary. The code is crude, but this _is_ just a test program. The
2462 default size has been increased from 32K to 50K.
2463
2464 2. The code in pcre_study() was using the value of the re argument before
2465 testing it for NULL. (Of course, in any sensible call of the function, it
2466 won't be NULL.)
2467
2468 3. The memmove() emulation function in pcre_internal.h, which is used on
2469 systems that lack both memmove() and bcopy() - that is, hardly ever -
2470 was missing a "static" storage class specifier.
2471
2472 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2473 containing an extended class (one that cannot be represented by a bitmap
2474 because it contains high-valued characters or Unicode property items, e.g.
2475 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2476 pattern, but PCRE should not loop if you do not (it no longer does).
2477 [Detail: two cases were found: (a) a repeated subpattern containing an
2478 extended class; (b) a recursive reference to a subpattern that followed a
2479 previous extended class. It wasn't skipping over the extended class
2480 correctly when UTF-8 mode was not set.]
2481
2482 5. A negated single-character class was not being recognized as fixed-length
2483 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2484 compile error "lookbehind assertion is not fixed length".
2485
2486 6. The RunPerlTest auxiliary script was showing an unexpected difference
2487 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2488 write a Perl script that can interpret lines of an input file either as
2489 byte characters or as UTF-8, which is what "perltest" was being required to
2490 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2491 can't do is switch easily at run time between having the "use utf8;" pragma
2492 or not. In the end, I fudged it by using the RunPerlTest script to insert
2493 "use utf8;" explicitly for the UTF-8 tests.
2494
2495 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2496 the end of the subject string, contrary to the documentation and to what
2497 Perl does. This was true of both matching functions. Now it matches only at
2498 the start of the subject and immediately after *internal* newlines.
2499
2500 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2501 a pointer to an int instead of a pointer to an unsigned long int. This
2502 caused problems on 64-bit systems.
2503
2504 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2505 instance of the 'standard' template library not being so standard".
2506
2507 10. There was no check on the number of named subpatterns nor the maximum
2508 length of a subpattern name. The product of these values is used to compute
2509 the size of the memory block for a compiled pattern. By supplying a very
2510 long subpattern name and a large number of named subpatterns, the size
2511 computation could be caused to overflow. This is now prevented by limiting
2512 the length of names to 32 characters, and the number of named subpatterns
2513 to 10,000.
2514
2515 11. Subpatterns that are repeated with specific counts have to be replicated in
2516 the compiled pattern. The size of memory for this was computed from the
2517 length of the subpattern and the repeat count. The latter is limited to
2518 65535, but there was no limit on the former, meaning that integer overflow
2519 could in principle occur. The compiled length of a repeated subpattern is
2520 now limited to 30,000 bytes in order to prevent this.
2521
2522 12. Added the optional facility to have named substrings with the same name.
2523
2524 13. Added the ability to use a named substring as a condition, using the
2525 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2526 are numbers (not recommended). Forward references are permitted.
2527
2528 14. Added forward references in named backreferences (if you see what I mean).
2529
2530 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2531 pattern could run off the end of the subject. For example, the pattern
2532 "(?s)(.{1,5})"8 did this with the subject "ab".
2533
2534 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2535 PCRE_CASELESS was set when matching characters that were quantified with ?
2536 or *.
2537
2538 17. A character class other than a single negated character that had a minimum
2539 but no maximum quantifier - for example [ab]{6,} - was not handled
2540 correctly by pce_dfa_exec(). It would match only one character.
2541
2542 18. A valid (though odd) pattern that looked like a POSIX character
2543 class but used an invalid character after [ (for example [[,abc,]]) caused
2544 pcre_compile() to give the error "Failed: internal error: code overflow" or
2545 in some cases to crash with a glibc free() error. This could even happen if
2546 the pattern terminated after [[ but there just happened to be a sequence of
2547 letters, a binary zero, and a closing ] in the memory that followed.
2548
2549 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2550 over the years. Originally (before any Unicode support), just the bottom 8
2551 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2552 output from "man perlunicode" includes this:
2553
2554 The regular expression compiler produces polymorphic opcodes. That
2555 is, the pattern adapts to the data and automatically switches to
2556 the Unicode character scheme when presented with Unicode data--or
2557 instead uses a traditional byte scheme when presented with byte
2558 data.
2559
2560 Sadly, a wide octal escape does not cause a switch, and in a string with
2561 no other multibyte characters, these octal escapes are treated as before.
2562 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2563 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2564 Unicode string.
2565
2566 I have not perpetrated such confusion in PCRE. Up till now, it took just
2567 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2568 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2569 translate to the appropriate multibyte character.
2570
2571 29. Applied some refactoring to reduce the number of warnings from Microsoft
2572 and Borland compilers. This has included removing the fudge introduced
2573 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2574 a warning about an unused variable.
2575
2576 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2577 characters since release 4.0, because Perl (from release 5.004) does not.
2578 [Or at least, is documented not to: some releases seem to be in conflict
2579 with the documentation.] However, when a pattern was studied with
2580 pcre_study() and all its branches started with \s, PCRE still included VT
2581 as a possible starting character. Of course, this did no harm; it just
2582 caused an unnecessary match attempt.
2583
2584 22. Removed a now-redundant internal flag bit that recorded the fact that case
2585 dependency changed within the pattern. This was once needed for "required
2586 byte" processing, but is no longer used. This recovers a now-scarce options
2587 bit. Also moved the least significant internal flag bit to the most-
2588 significant bit of the word, which was not previously used (hangover from
2589 the days when it was an int rather than a uint) to free up another bit for
2590 the future.
2591
2592 23. Added support for CRLF line endings as well as CR and LF. As well as the
2593 default being selectable at build time, it can now be changed at runtime
2594 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2595 specify that it is scanning data with non-default line endings.
2596
2597 24. Changed the definition of CXXLINK to make it agree with the definition of
2598 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2599
2600 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2601 recursions. This makes a big different to stack usage for some patterns.
2602
2603 26. If a subpattern containing a named recursion or subroutine reference such
2604 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2605 the space required for the compiled pattern went wrong and gave too small a
2606 value. Depending on the environment, this could lead to "Failed: internal
2607 error: code overflow at offset 49" or "glibc detected double free or
2608 corruption" errors.
2609
2610 27. Applied patches from Google (a) to support the new newline modes and (b) to
2611 advance over multibyte UTF-8 characters in GlobalReplace.
2612
2613 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2614 difference for some implementation of PCRE in some Windows version.
2615
2616 29. Added some extra testing facilities to pcretest:
2617
2618 \q<number> in a data line sets the "match limit" value
2619 \Q<number> in a data line sets the "match recursion limt" value
2620 -S <number> sets the stack size, where <number> is in megabytes
2621
2622 The -S option isn't available for Windows.
2623
2624
2625 Version 6.6 06-Feb-06
2626 ---------------------
2627
2628 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2629 in pcreposix.h. I have copied the definition from pcre.h.
2630
2631 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2632 because pcre.h is no longer a built file.
2633
2634 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2635 not normally included in the compiled code.
2636
2637
2638 Version 6.5 01-Feb-06
2639 ---------------------
2640
2641 1. When using the partial match feature with pcre_dfa_exec(), it was not
2642 anchoring the second and subsequent partial matches at the new starting
2643 point. This could lead to incorrect results. For example, with the pattern
2644 /1234/, partially matching against "123" and then "a4" gave a match.
2645
2646 2. Changes to pcregrep:
2647
2648 (a) All non-match returns from pcre_exec() were being treated as failures
2649 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2650 error message is output. Some extra information is given for the
2651 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2652 probably the only errors that are likely to be caused by users (by
2653 specifying a regex that has nested indefinite repeats, for instance).
2654 If there are more than 20 of these errors, pcregrep is abandoned.
2655
2656 (b) A binary zero was treated as data while matching, but terminated the
2657 output line if it was written out. This has been fixed: binary zeroes
2658 are now no different to any other data bytes.
2659
2660 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2661 used to set a locale for matching. The --locale=xxxx long option has
2662 been added (no short equivalent) to specify a locale explicitly on the
2663 pcregrep command, overriding the environment variables.
2664
2665 (d) When -B was used with -n, some line numbers in the output were one less
2666 than they should have been.
2667
2668 (e) Added the -o (--only-matching) option.
2669
2670 (f) If -A or -C was used with -c (count only), some lines of context were
2671 accidentally printed for the final match.
2672
2673 (g) Added the -H (--with-filename) option.
2674
2675 (h) The combination of options -rh failed to suppress file names for files
2676 that were found from directory arguments.
2677
2678 (i) Added the -D (--devices) and -d (--directories) options.
2679
2680 (j) Added the -F (--fixed-strings) option.
2681
2682 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2683
2684 (l) Added the --colo(u)r option.
2685
2686 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2687 is not present by default.
2688
2689 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2690 items such as (?R) or (?1), when the recursion could match a number of
2691 alternatives. If it matched one of the alternatives, but subsequently,
2692 outside the recursion, there was a failure, the code tried to back up into
2693 the recursion. However, because of the way PCRE is implemented, this is not
2694 possible, and the result was an incorrect result from the match.
2695
2696 In order to prevent this happening, the specification of recursion has
2697 been changed so that all such subpatterns are automatically treated as
2698 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2699
2700 4. I had overlooked the fact that, in some locales, there are characters for
2701 which isalpha() is true but neither isupper() nor islower() are true. In
2702 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2703 and ordfeminine) are like this. This affected the treatment of \w and \W
2704 when they appeared in character classes, but not when they appeared outside
2705 a character class. The bit map for "word" characters is now created
2706 separately from the results of isalnum() instead of just taking it from the
2707 upper, lower, and digit maps. (Plus the underscore character, of course.)
2708
2709 5. The above bug also affected the handling of POSIX character classes such as
2710 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2711 permanent tables. Instead, the bit maps for such a class were previously
2712 created as the appropriate unions of the upper, lower, and digit bitmaps.
2713 Now they are created by subtraction from the [[:word:]] class, which has
2714 its own bitmap.
2715
2716 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2717 It is created by subtracting the vertical space characters (\x09, \x0a,
2718 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2719 subtraction was done in the overall bitmap for a character class, meaning
2720 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2721 be recognized. This bug has been fixed.
2722
2723 7. Patches from the folks at Google:
2724
2725 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2726 real life, but is still worth protecting against".
2727
2728 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2729 regular expressions".
2730
2731 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2732 have it.
2733
2734 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2735 "configure" and the latter not, in order to fix a problem somebody had
2736 with compiling the Arg class on HP-UX.
2737
2738 (e) Improve the error-handling of the C++ wrapper a little bit.
2739
2740 (f) New tests for checking recursion limiting.
2741
2742 8. The pcre_memmove() function, which is used only if the environment does not
2743 have a standard memmove() function (and is therefore rarely compiled),
2744 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2745 returning a result (though PCRE never actually uses the result).
2746
2747 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2748 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2749 returned instead of calling malloc() with an overflowing number that would
2750 most likely cause subsequent chaos.
2751
2752 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2753
2754 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2755 with this option is matched, the nmatch and pmatch options of regexec() are
2756 ignored.
2757
2758 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2759 provided in case anyone wants to the the POSIX interface with UTF-8
2760 strings.
2761
2762 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2763 C++ linking (needed for some HP-UX environments).
2764
2765 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2766 (unused parameter) and in the pcre_printint() function (omitted "default"
2767 switch label when the default is to do nothing).
2768
2769 15. Added some code to make it possible, when PCRE is compiled as a C++
2770 library, to replace subject pointers for pcre_exec() with a smart pointer
2771 class, thus making it possible to process discontinuous strings.
2772
2773 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2774 much the same function. They were added by different people who were trying
2775 to make PCRE easy to compile on non-Unix systems. It has been suggested
2776 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2777 for compiling on Windows systems. I have therefore replaced it with
2778 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2779 defaults to "extern" for C or "extern C" for C++, which works fine on
2780 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2781 SCOPE with something explicit in config.h. In addition:
2782
2783 (a) pcreposix.h still had just "extern" instead of either of these macros;
2784 I have replaced it with PCRE_DATA_SCOPE.
2785
2786 (b) Functions such as _pcre_xclass(), which are internal to the library,
2787 but external in the C sense, all had PCRE_EXPORT in their definitions.
2788 This is apparently wrong for the Windows case, so I have removed it.
2789 (It makes no difference on Unix-like systems.)
2790
2791 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2792 of recursive calls to match(). This is different to MATCH_LIMIT because
2793 that limits the total number of calls to match(), not all of which increase
2794 the depth of recursion. Limiting the recursion depth limits the amount of
2795 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2796 when PCRE is compiled, and changed at run time. A patch from Google adds
2797 this functionality to the C++ interface.
2798
2799 18. Changes to the handling of Unicode character properties:
2800
2801 (a) Updated the table to Unicode 4.1.0.
2802
2803 (b) Recognize characters that are not in the table as "Cn" (undefined).
2804
2805 (c) I revised the way the table is implemented to a much improved format
2806 which includes recognition of ranges. It now supports the ranges that
2807 are defined in UnicodeData.txt, and it also amalgamates other
2808 characters into ranges. This has reduced the number of entries in the
2809 table from around 16,000 to around 3,000, thus reducing its size
2810 considerably. I realized I did not need to use a tree structure after
2811 all - a binary chop search is just as efficient. Having reduced the
2812 number of entries, I extended their size from 6 bytes to 8 bytes to
2813 allow for more data.
2814
2815 (d) Added support for Unicode script names via properties such as \p{Han}.
2816
2817 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2818 matching that character.
2819
2820 20. When matching a repeated Unicode property with a minimum greater than zero,
2821 (for example \pL{2,}), PCRE could look past the end of the subject if it
2822 reached it while seeking the minimum number of characters. This could
2823 happen only if some of the characters were more than one byte long, because
2824 there is a check for at least the minimum number of bytes.
2825
2826 21. Refactored the implementation of \p and \P so as to be more general, to
2827 allow for more different types of property in future. This has changed the
2828 compiled form incompatibly. Anybody with saved compiled patterns that use
2829 \p or \P will have to recompile them.
2830
2831 22. Added "Any" and "L&" to the supported property types.
2832
2833 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2834 but give a compile time error if the value is greater than 0xff.
2835
2836 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2837 accidentally not being installed or uninstalled.
2838
2839 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2840 made were to insert the current release number. This seemed silly, because
2841 it made things harder for people building PCRE on systems that don't run
2842 "configure". I have turned pcre.h into a distributed file, no longer built
2843 by "configure", with the version identification directly included. There is
2844 no longer a pcre.h.in file.
2845
2846 However, this change necessitated a change to the pcre-config script as
2847 well. It is built from pcre-config.in, and one of the substitutions was the
2848 release number. I have updated configure.ac so that ./configure now finds
2849 the release number by grepping pcre.h.
2850
2851 26. Added the ability to run the tests under valgrind.
2852
2853
2854 Version 6.4 05-Sep-05
2855 ---------------------
2856
2857 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2858 "--" to be printed when multiple files were scanned, even when none of the
2859 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2860 consider it to be a bug, and have restored the previous behaviour.
2861
2862 2. A couple of code tidies to get rid of compiler warnings.
2863
2864 3. The pcretest program used to cheat by referring to symbols in the library
2865 whose names begin with _pcre_. These are internal symbols that are not
2866 really supposed to be visible externally, and in some environments it is
2867 possible to suppress them. The cheating is now confined to including
2868 certain files from the library's source, which is a bit cleaner.
2869
2870 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2871 file's purpose clearer.
2872
2873 5. Reorganized pcre_ucp_findchar().
2874
2875
2876 Version 6.3 15-Aug-05
2877 ---------------------
2878
2879 1. The file libpcre.pc.in did not have general read permission in the tarball.
2880
2881 2. There were some problems when building without C++ support:
2882
2883 (a) If C++ support was not built, "make install" and "make test" still
2884 tried to test it.
2885
2886 (b) There were problems when the value of CXX was explicitly set. Some
2887 changes have been made to try to fix these, and ...
2888
2889 (c) --disable-cpp can now be used to explicitly disable C++ support.
2890
2891 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2892 backslash in a target when C++ was disabled. This confuses some
2893 versions of "make", apparently. Using an intermediate variable solves
2894 this. (Same for CPP_LOBJ.)
2895
2896 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2897 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2898 necessary on certain architectures.
2899
2900 4. Added a setting of -export-symbols-regex to the link command to remove
2901 those symbols that are exported in the C sense, but actually are local
2902 within the library, and not documented. Their names all begin with
2903 "_pcre_". This is not a perfect job, because (a) we have to except some
2904 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2905 available (and never for static libraries). I have made a note to try to
2906 find a way round (a) in the future.
2907
2908
2909 Version 6.2 01-Aug-05
2910 ---------------------
2911
2912 1. There was no test for integer overflow of quantifier values. A construction
2913 such as {1111111111111111} would give undefined results. What is worse, if
2914 a minimum quantifier for a parenthesized subpattern overflowed and became
2915 negative, the calculation of the memory size went wrong. This could have
2916 led to memory overwriting.
2917
2918 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2919
2920 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2921 operating environments where this matters.
2922
2923 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2924 PCRE options from within the C++ wrapper.
2925
2926 5. Named capturing subpatterns were not being correctly counted when a pattern
2927 was compiled. This caused two problems: (a) If there were more than 100
2928 such subpatterns, the calculation of the memory needed for the whole
2929 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2930 back references of the form \12, where the number was greater than 9, were
2931 not recognized as back references, even though there were sufficient
2932 previous subpatterns.
2933
2934 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2935 versions of gcc, e.g. 2.95.4.
2936
2937
2938 Version 6.1 21-Jun-05
2939 ---------------------
2940
2941 1. There was one reference to the variable "posix" in pcretest.c that was not
2942 surrounded by "#if !defined NOPOSIX".
2943
2944 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2945 the cross-check on the old pcre_info() function, for the benefit of the
2946 cut-down version of PCRE that is currently imported into Exim.
2947
2948 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2949 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2950 patterns that start (?i) but I don't think that matters. The use of (?i) is
2951 just an example; this all applies to the other options as well.
2952
2953 4. Since libtool seems to echo the compile commands it is issuing, the output
2954 from "make" can be reduced a bit by putting "@" in front of each libtool
2955 compile command.
2956
2957 5. Patch from the folks at Google for configure.in to be a bit more thorough
2958 in checking for a suitable C++ installation before trying to compile the
2959 C++ stuff. This should fix a reported problem when a compiler was present,
2960 but no suitable headers.
2961
2962 6. The man pages all had just "PCRE" as their title. I have changed them to
2963 be the relevant file name. I have also arranged that these names are
2964 retained in the file doc/pcre.txt, which is a concatenation in text format
2965 of all the man pages except the little individual ones for each function.
2966
2967 7. The NON-UNIX-USE file had not been updated for the different set of source
2968 files that come with release 6. I also added a few comments about the C++
2969 wrapper.
2970
2971
2972 Version 6.0 07-Jun-05
2973 ---------------------
2974
2975 1. Some minor internal re-organization to help with my DFA experiments.
2976
2977 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2978 didn't matter for the library itself when fully configured, but did matter
2979 when compiling without UCP support, or within Exim, where the ucp files are
2980 not imported.
2981
2982 3. Refactoring of the library code to split up the various functions into
2983 different source modules. The addition of the new DFA matching code (see
2984 below) to a single monolithic source would have made it really too
2985 unwieldy, quite apart from causing all the code to be include in a
2986 statically linked application, when only some functions are used. This is
2987 relevant even without the DFA addition now that patterns can be compiled in
2988 one application and matched in another.
2989
2990 The downside of splitting up is that there have to be some external
2991 functions and data tables that are used internally in different modules of
2992 the library but which are not part of the API. These have all had their
2993 names changed to start with "_pcre_" so that they are unlikely to clash
2994 with other external names.
2995
2996 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2997 a different (DFA) algorithm. Although it is slower than the original
2998 function, it does have some advantages for certain types of matching
2999 problem.
3000
3001 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3002 including restarting after a partial match.
3003
3004 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3005 defined when compiling for Windows was sent to me. I have put it into the
3006 code, though I have no means of testing or verifying it.
3007
3008 7. Added the pcre_refcount() auxiliary function.
3009
3010 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3011 match before or at the first newline in the subject string. In pcretest,
3012 the /f option on a pattern can be used to set this.
3013
3014 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3015 would behave wrongly. This has been present in PCRE since release 4.0.
3016
3017 10. A number of changes to the pcregrep command:
3018
3019 (a) Refactored how -x works; insert ^(...)$ instead of setting
3020 PCRE_ANCHORED and checking the length, in preparation for adding
3021 something similar for -w.
3022
3023 (b) Added the -w (match as a word) option.
3024
3025 (c) Refactored the way lines are read and buffered so as to have more
3026 than one at a time available.
3027
3028 (d) Implemented a pcregrep test script.
3029
3030 (e) Added the -M (multiline match) option. This allows patterns to match
3031 over several lines of the subject. The buffering ensures that at least
3032 8K, or the rest of the document (whichever is the shorter) is available
3033 for matching (and similarly the previous 8K for lookbehind assertions).
3034
3035 (f) Changed the --help output so that it now says
3036
3037 -w, --word-regex(p)
3038
3039 instead of two lines, one with "regex" and the other with "regexp"
3040 because that confused at least one person since the short forms are the
3041 same. (This required a bit of code, as the output is generated
3042 automatically from a table. It wasn't just a text change.)
3043
3044 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3045 option but starts with a hyphen. Could be a pattern or a path name
3046 starting with a hyphen, for instance.
3047
3048 (h) "-" can be given as a file name to represent stdin.
3049
3050 (i) When file names are being printed, "(standard input)" is used for
3051 the standard input, for compatibility with GNU grep. Previously
3052 "<stdin>" was used.
3053
3054 (j) The option --label=xxx can be used to supply a name to be used for
3055 stdin when file names are being printed. There is no short form.
3056
3057 (k) Re-factored the options decoding logic because we are going to add
3058 two more options that take data. Such options can now be given in four
3059 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3060
3061 (l) Added the -A, -B, and -C options for requesting that lines of context
3062 around matches be printed.
3063
3064 (m) Added the -L option to print the names of files that do not contain
3065 any matching lines, that is, the complement of -l.
3066
3067 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3068 continue to scan other files.
3069
3070 (o) The -s option was incorrectly implemented. For compatibility with other
3071 greps, it now suppresses the error message for a non-existent or non-
3072 accessible file (but not the return code). There is a new option called
3073 -q that suppresses the output of matching lines, which was what -s was
3074 previously doing.
3075
3076 (p) Added --include and --exclude options to specify files for inclusion
3077 and exclusion when recursing.
3078
3079 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3080 Hopefully, it now does.
3081
3082 12. Missing cast in pcre_study().
3083
3084 13. Added an "uninstall" target to the makefile.
3085
3086 14. Replaced "extern" in the function prototypes in Makefile.in with
3087 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3088 world, but is set differently for Windows.
3089
3090 15. Added a second compiling function called pcre_compile2(). The only
3091 difference is that it has an extra argument, which is a pointer to an
3092 integer error code. When there is a compile-time failure, this is set
3093 non-zero, in addition to the error test pointer being set to point to an
3094 error message. The new argument may be NULL if no error number is required
3095 (but then you may as well call pcre_compile(), which is now just a
3096 wrapper). This facility is provided because some applications need a
3097 numeric error indication, but it has also enabled me to tidy up the way
3098 compile-time errors are handled in the POSIX wrapper.
3099
3100 16. Added VPATH=.libs to the makefile; this should help when building with one
3101 prefix path and installing with another. (Or so I'm told by someone who
3102 knows more about this stuff than I do.)
3103
3104 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3105 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3106 match everything, including newlines. This is not POSIX-compatible, but
3107 somebody wanted the feature. From pcretest it can be activated by using
3108 both the P and the s flags.
3109
3110 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3111
3112 19. libpcre.pc was being incorrectly installed as executable.
3113
3114 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3115 it now also looks for '\r' so that it will work unmodified on Windows.
3116
3117 21. Added Google's contributed C++ wrapper to the distribution.
3118
3119 22. Added some untidy missing memory free() calls in pcretest, to keep
3120 Electric Fence happy when testing.
3121
3122
3123
3124 Version 5.0 13-Sep-04
3125 ---------------------
3126
3127 1. Internal change: literal characters are no longer packed up into items
3128 containing multiple characters in a single byte-string. Each character
3129 is now matched using a separate opcode. However, there may be more than one
3130 byte in the character in UTF-8 mode.
3131
3132 2. The pcre_callout_block structure has two new fields: pattern_position and
3133 next_item_length. These contain the offset in the pattern to the next match
3134 item, and its length, respectively.
3135
3136 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3137 insertion of callouts before each pattern item. Added the /C option to
3138 pcretest to make use of this.
3139
3140 4. On the advice of a Windows user, the lines
3141
3142 #if defined(_WIN32) || defined(WIN32)
3143 _setmode( _fileno( stdout ), 0x8000 );
3144 #endif /* defined(_WIN32) || defined(WIN32) */
3145
3146 have been added to the source of pcretest. This apparently does useful
3147 magic in relation to line terminators.
3148
3149 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3150 for the benefit of those environments where the "b" makes a difference.
3151
3152 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3153 to know about it. I have put a hack into configure.in that adds in code
3154 to set GCC=yes if CC=icc. This seems to end up at a point in the
3155 generated configure script that is early enough to affect the setting of
3156 compiler options, which is what is needed, but I have no means of testing
3157 whether it really works. (The user who reported this had patched the
3158 generated configure script, which of course I cannot do.)
3159
3160 LATER: After change 22 below (new libtool files), the configure script
3161 seems to know about icc (and also ecc). Therefore, I have commented out
3162 this hack in configure.in.
3163
3164 7. Added support for pkg-config (2 patches were sent in).
3165
3166 8. Negated POSIX character classes that used a combination of internal tables
3167 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3168 [[:^ascii]]. Typically, they would match almost any characters. The other
3169 POSIX classes were not broken in this way.
3170
3171 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3172 to find the match, as PCRE was deluded into thinking that the match had to
3173 start at the start point or following a newline. The same bug applied to
3174 patterns with negative forward assertions or any backward assertions
3175 preceding ".*" at the start, unless the pattern required a fixed first
3176 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3177
3178 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3179 starting at the last subject character, bytes beyond the end of the subject
3180 string were read.
3181
3182 11. Renamed the variable "class" as "classbits" to make life easier for C++
3183 users. (Previously there was a macro definition, but it apparently wasn't
3184 enough.)
3185
3186 12. Added the new field "tables" to the extra data so that tables can be passed
3187 in at exec time, or the internal tables can be re-selected. This allows
3188 a compiled regex to be saved and re-used at a later time by a different
3189 program that might have everything at different addresses.
3190
3191 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3192 -R library as well as a -L library.
3193
3194 14. The debugging options of pcretest (-d on the command line or D on a
3195 pattern) showed incorrect output for anything following an extended class
3196 that contained multibyte characters and which was followed by a quantifier.
3197
3198 15. Added optional support for general category Unicode character properties
3199 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3200 support. It adds about 90K to the size of the library. The meanings of the
3201 inbuilt class escapes such as \d and \s have NOT been changed.
3202
3203 16. Updated pcredemo.c to include calls to free() to release the memory for the
3204 compiled pattern.
3205
3206 17. The generated file chartables.c was being created in the source directory
3207 instead of in the building directory. This caused the build to fail if the
3208 source directory was different from the building directory, and was
3209 read-only.
3210
3211 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3212 file. No doubt somebody will tell me if they don't make sense... Also added
3213 Dan Mooney's comments about building on OpenVMS.
3214
3215 19. Added support for partial matching via the PCRE_PARTIAL option for
3216 pcre_exec() and the \P data escape in pcretest.
3217
3218 20. Extended pcretest with 3 new pattern features:
3219
3220 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3221 write the compiled pattern to the file whose name is "rest-of-line".
3222 This is a straight binary dump of the data, with the saved pointer to
3223 the character tables forced to be NULL. The study data, if any, is
3224 written too. After writing, pcretest reads a new pattern.
3225
3226 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3227 compiled pattern from the given file. There must not be any
3228 occurrences of "<" in the file name (pretty unlikely); if there are,
3229 pcretest will instead treat the initial "<" as a pattern delimiter.
3230 After reading in the pattern, pcretest goes on to read data lines as
3231 usual.
3232
3233 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3234 and 16-bit fields in a compiled pattern, to simulate a pattern that
3235 was compiled on a host of opposite endianness.
3236
3237 21. The pcre-exec() function can now cope with patterns that were compiled on
3238 hosts of opposite endianness, with this restriction:
3239
3240 As for any compiled expression that is saved and used later, the tables
3241 pointer field cannot be preserved; the extra_data field in the arguments
3242 to pcre_exec() should be used to pass in a tables address if a value
3243 other than the default internal tables were used at compile time.
3244
3245 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3246 now diagnosed as an error. Previously, most of the time, a negative number
3247 would have been treated as zero, but if in addition "ovector" was passed as
3248 NULL, a crash could occur.
3249
3250 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3251 new versions from the libtool 1.5 distribution (the last one is a copy of
3252 a file called libtool.m4). This seems to have fixed the need to patch
3253 "configure" to support Darwin 1.3 (which I used to do). However, I still
3254 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3255 workstation).
3256
3257 24. Changed the PCRE licence to be the more standard "BSD" licence.
3258
3259
3260 Version 4.5 01-Dec-03
3261 ---------------------
3262
3263 1. There has been some re-arrangement of the code for the match() function so
3264 that it can be compiled in a version that does not call itself recursively.
3265 Instead, it keeps those local variables that need separate instances for
3266 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3267 needs to "recurse". Keeping track of where control must go is done by means
3268 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3269 hide most of the details from the main code, and operates only if
3270 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3271 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3272 operating.
3273
3274 To make it easier for callers to provide specially tailored get/free
3275 functions for this usage, two new functions, pcre_stack_malloc, and
3276 pcre_stack_free, are used. They are always called in strict stacking order,
3277 and the size of block requested is always the same.
3278
3279 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3280 PCRE has been compiled to use the stack or the heap for recursion. The
3281 -C option of pcretest uses this to show which version is compiled.
3282
3283 A new data escape \S, is added to pcretest; it causes the amounts of store
3284 obtained and freed by both kinds of malloc/free at match time to be added
3285 to the output.
3286
3287 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3288 what's available on my current Linux desktop machine.
3289
3290 3. When matching a UTF-8 string, the test for a valid string at the start has
3291 been extended. If start_offset is not zero, PCRE now checks that it points
3292 to a byte that is the start of a UTF-8 character. If not, it returns
3293 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3294 this is necessary because there may be backward assertions in the pattern.
3295 When matching the same subject several times, it may save resources to use
3296 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3297
3298 4. The code for checking the validity of UTF-8 strings has been tightened so
3299 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3300 containing "overlong sequences".
3301
3302 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3303 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3304 should have been "&", but it just so happened that all the cases this let
3305 through by mistake were picked up later in the function.
3306
3307 6. I had used a variable called "isblank" - this is a C99 function, causing
3308 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3309
3310 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3311 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3312 so the version is shown; (c) stop "make test" echoing "./RunTest".
3313
3314 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3315
3316 9. The prototype for memmove() for systems that don't have it was using
3317 size_t, but the inclusion of the header that defines size_t was later. I've
3318 moved the #includes for the C headers earlier to avoid this.
3319
3320 10. Added some adjustments to the code to make it easier to compiler on certain
3321 special systems:
3322
3323 (a) Some "const" qualifiers were missing.
3324 (b) Added the macro EXPORT before all exported functions; by default this
3325 is defined to be empty.
3326 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3327 that it reads its output file name as an argument instead of writing
3328 to the standard output and assuming this can be redirected.
3329
3330 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3331 class containing characters with values greater than 255, PCRE compilation
3332 went into a loop.
3333
3334 12. A recursive reference to a subpattern that was within another subpattern
3335 that had a minimum quantifier of zero caused PCRE to crash. For example,
3336 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3337 recursion. If the recursively-called subpattern itself had a zero repeat,
3338 that was OK.
3339
3340 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3341 buffer into which it was copied (for escape processing) was still set at
3342 1024, so long lines caused crashes.
3343
3344 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3345 "internal error: code overflow...". This applied to any character class
3346 that was followed by a possessive quantifier.
3347
3348 15. Modified the Makefile to add libpcre.la as a prerequisite for
3349 libpcreposix.la because I was told this is needed for a parallel build to
3350 work.
3351
3352 16. If a pattern that contained .* following optional items at the start was
3353 studied, the wrong optimizing data was generated, leading to matching
3354 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3355 matching string must start with a or b or c. The correct conclusion for
3356 this pattern is that a match can start with any character.
3357
3358
3359 Version 4.4 13-Aug-03
3360 ---------------------
3361
3362 1. In UTF-8 mode, a character class containing characters with values between
3363 127 and 255 was not handled correctly if the compiled pattern was studied.
3364 In fixing this, I have also improved the studying algorithm for such
3365 classes (slightly).
3366
3367 2. Three internal functions had redundant arguments passed to them. Removal
3368 might give a very teeny performance improvement.
3369
3370 3. Documentation bug: the value of the capture_top field in a callout is *one
3371 more than* the number of the hightest numbered captured substring.
3372
3373 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3374 in incorrectly linking with a previously installed version. They now link
3375 explicitly with libpcre.la.
3376
3377 5. configure.in no longer needs to recognize Cygwin specially.
3378
3379 6. A problem in pcre.in for Windows platforms is fixed.
3380
3381 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3382 pcretest, it used to include the size of the study block as part of its
3383 output. Unfortunately, the structure contains a field that has a different
3384 size on different hardware architectures. This meant that the tests that
3385 showed this size failed. As the block is currently always of a fixed size,
3386 this information isn't actually particularly useful in pcretest output, so
3387 I have just removed it.
3388
3389 8. Three pre-processor statements accidentally did not start in column 1.
3390 Sadly, there are *still* compilers around that complain, even though
3391 standard C has not required this for well over a decade. Sigh.
3392
3393 9. In pcretest, the code for checking callouts passed small integers in the
3394 callout_data field, which is a void * field. However, some picky compilers
3395 complained about the casts involved for this on 64-bit systems. Now
3396 pcretest passes the address of the small integer instead, which should get
3397 rid of the warnings.
3398
3399 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3400 both compile and run time, and gives an error if an invalid UTF-8 sequence
3401 is found. There is a option for disabling this check in cases where the
3402 string is known to be correct and/or the maximum performance is wanted.
3403
3404 11. In response to a bug report, I changed one line in Makefile.in from
3405
3406 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3407 to
3408 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3409
3410 to look similar to other lines, but I have no way of telling whether this
3411 is the right thing to do, as I do not use Windows. No doubt I'll get told
3412 if it's wrong...
3413
3414
3415 Version 4.3 21-May-03
3416 ---------------------
3417
3418 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3419 Makefile.
3420
3421 2. Some refactoring to improve the quality of the code:
3422
3423 (i) The utf8_table... variables are now declared "const".
3424
3425 (ii) The code for \cx, which used the "case flipping" table to upper case
3426 lower case letters, now just substracts 32. This is ASCII-specific,
3427 but the whole concept of \cx is ASCII-specific, so it seems
3428 reasonable.
3429
3430 (iii) PCRE was using its character types table to recognize decimal and
3431 hexadecimal digits in the pattern. This is silly, because it handles
3432 only 0-9, a-f, and A-F, but the character types table is locale-
3433 specific, which means strange things might happen. A private
3434 table is now used for this - though it costs 256 bytes, a table is
3435 much faster than multiple explicit tests. Of course, the standard
3436 character types table is still used for matching digits in subject
3437 strings against \d.
3438
3439 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3440 ending in _t are). So I've renamed it as ESC_tee.
3441
3442 3. The first argument for regexec() in the POSIX wrapper should have been
3443 defined as "const".
3444
3445 4. Changed pcretest to use malloc() for its buffers so that they can be
3446 Electric Fenced for debugging.
3447
3448 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3449 to read one or more bytes before the start of the subject string. Often this
3450 had no effect on PCRE's behaviour, but in some circumstances it could
3451 provoke a segmentation fault.
3452
3453 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3454 to try to read one or more bytes before the start of the subject string.
3455
3456 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3457 UTF-8 support could misbehave in various ways if the subject string
3458 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3459 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3460 back over UTF-8 characters.)
3461
3462
3463 Version 4.2 14-Apr-03
3464 ---------------------
3465
3466 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3467
3468 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3469 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3470 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3471 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3472 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3473 and BUILD_EXEEXT
3474 Note: automatic setting of the BUILD variables is not yet working
3475 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3476 compile-time but not at link-time
3477 [LINK]: use for linking executables only
3478 make different versions for Windows and non-Windows
3479 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3480 libraries
3481 [LINK_FOR_BUILD]: new variable
3482 [OBJEXT]: use throughout
3483 [EXEEXT]: use throughout
3484 <winshared>: new target
3485 <wininstall>: new target
3486 <dftables.o>: use native compiler
3487 <dftables>: use native linker
3488 <install>: handle Windows platform correctly
3489 <clean>: ditto
3490 <check>: ditto
3491 copy DLL to top builddir before testing
3492
3493 As part of these changes, -no-undefined was removed again. This was reported
3494 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3495 in any case.
3496
3497 3. Some tidies to get rid of compiler warnings:
3498
3499 . In the match_data structure, match_limit was an unsigned long int, whereas
3500 match_call_count was an int. I've made them both unsigned long ints.
3501
3502 . In pcretest the fact that a const uschar * doesn't automatically cast to
3503 a void * provoked a warning.
3504
3505 . Turning on some more compiler warnings threw up some "shadow" variables
3506 and a few more missing casts.
3507
3508 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3509 option, a class that contained a single character with a value between 128
3510 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3511
3512 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3513 option, a class that contained several characters, but with at least one
3514 whose value was between 128 and 255 caused PCRE to crash.
3515
3516
3517 Version 4.1 12-Mar-03
3518 ---------------------
3519
3520 1. Compiling with gcc -pedantic found a couple of places where casts were
3521 needed, and a string in dftables.c that was longer than standard compilers are
3522 required to support.
3523
3524 2. Compiling with Sun's compiler found a few more places where the code could
3525 be tidied up in order to avoid warnings.
3526
3527 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3528 first of these names is deprecated in the latest Autoconf in favour of the name
3529 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3530 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3531 analogy I have changed it to CFLAGS_FOR_BUILD.
3532
3533 4. Added -no-undefined to the linking command in the Makefile, because this is
3534 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3535 linking step for the pcreposix library.
3536
3537 5. PCRE was failing to diagnose the case of two named groups with the same
3538 name.
3539
3540 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3541 literal character that is needed in the subject for a match, and scans along to
3542 ensure that it is present before embarking on the full matching process. This
3543 saves time in cases of nested unlimited repeats that are never going to match.
3544 Problem: the scan can take a lot of time if the subject is very long (e.g.
3545 megabytes), thus penalizing straightforward matches. It is now done only if the
3546 amount of subject to be scanned is less than 1000 bytes.
3547
3548 7. A lesser problem with the same optimization is that it was recording the
3549 first character of an anchored pattern as "needed", thus provoking a search
3550 right along the subject, even when the first match of the pattern was going to
3551 fail. The "needed" character is now not set for anchored patterns, unless it
3552 follows something in the pattern that is of non-fixed length. Thus, it still
3553 fulfils its original purpose of finding quick non-matches in cases of nested
3554 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3555
3556
3557 Version 4.0 17-Feb-03
3558 ---------------------
3559
3560 1. If a comment in an extended regex that started immediately after a meta-item
3561 extended to the end of string, PCRE compiled incorrect data. This could lead to
3562 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3563
3564 2. Moved to autoconf 2.53 and libtool 1.4.2.
3565
3566 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3567 the special perltest8 script is no longer needed - all the tests can be run
3568 from a single perltest script.
3569
3570 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3571 by \s. It has now been removed in PCRE. This means it isn't recognized as
3572 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3573 class [:space:] *does* include VT, thereby creating a mess.
3574
3575 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3576 space and tab.
3577
3578 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3579 its new features into the main test script, reducing the number of scripts.
3580
3581 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3582 were backward compatible, and made the (?i) apply to the whole pattern, as if
3583 /i were given. Now it behaves more logically, and applies the option setting
3584 only to what follows. PCRE has been changed to follow suit. However, if it
3585 finds options settings right at the start of the pattern, it extracts them into
3586 the global options, as before. Thus, they show up in the info data.
3587
3588 8. Added support for the \Q...\E escape sequence. Characters in between are
3589 treated as literals. This is slightly different from Perl in that $ and @ are
3590 also handled as literals inside the quotes. In Perl, they will cause variable
3591 interpolation. Note the following examples:
3592
3593 Pattern PCRE matches Perl matches
3594
3595 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3596 \Qabc\$xyz\E abc\$xyz abc\$xyz
3597 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3598
3599 For compatibility with Perl, \Q...\E sequences are recognized inside character
3600 classes as well as outside them.
3601
3602 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3603 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3604 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3605 signed/unsigned warnings.
3606
3607 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3608 option for pcretest, so I've replaced it by a simple function that does just
3609 that job.
3610
3611 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3612 "pcregrep -".
3613
3614 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3615 Java package. This provides some syntactic sugar for simple cases of what my
3616 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3617 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3618 item, you can use this simplified notation. Note that only makes sense with
3619 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3620 greediness, whatever the setting of the PCRE_UNGREEDY option.
3621
3622 13. A change of greediness default within a pattern was not taking effect at
3623 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3624 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3625 was abstracted outside.
3626
3627 14. PCRE now supports the \G assertion. It is true when the current matching
3628 position is at the start point of the match. This differs from \A when the
3629 starting offset is non-zero. Used with the /g option of pcretest (or similar
3630 code), it works in the same way as it does for Perl's /g option. If all
3631 alternatives of a regex begin with \G, the expression is anchored to the start
3632 match position, and the "anchored" flag is set in the compiled expression.
3633
3634 15. Some bugs concerning the handling of certain option changes within patterns
3635 have been fixed. These applied to options other than (?ims). For example,
3636 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3637 been the other way round. Some of this was related to change 7 above.
3638
3639 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3640 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3641 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3642 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3643
3644 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3645 mode. Unlike ".", it always matches newline, whatever the setting of
3646 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3647 assertions. Perl allows it, but it doesn't (in general) work because it can't
3648 calculate the length of the lookbehind. At least, that's the case for Perl
3649 5.8.0 - I've been told they are going to document that it doesn't work in
3650 future.
3651
3652 18. Added an error diagnosis for escapes that PCRE does not support: these are
3653 \L, \l, \N, \P, \p, \U, \u, and \X.
3654
3655 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3656 reading past the end of the pattern in cases such as /[abcd/.
3657
3658 20. PCRE was getting more memory than necessary for patterns with classes that
3659 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3660
3661 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3662 compiling PCRE for use with Virtual Pascal.
3663
3664 22. Small fix to the Makefile to make it work properly if the build is done
3665 outside the source tree.
3666
3667 23. Added a new extension: a condition to go with recursion. If a conditional
3668 subpattern starts with (?(R) the "true" branch is used if recursion has
3669 happened, whereas the "false" branch is used only at the top level.
3670
3671 24. When there was a very long string of literal characters (over 255 bytes
3672 without UTF support, over 250 bytes with UTF support), the computation of how
3673 much memory was required could be incorrect, leading to segfaults or other
3674 strange effects.
3675
3676 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3677 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3678 there was a subsequent back reference to those brackets. This meant that, for
3679 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3680 possible to check for precisely this case. All we can do is abandon the
3681 optimization if .* occurs inside capturing brackets when there are any back
3682 references whatsoever. (See below for a better fix that came later.)
3683
3684 26. The handling of the optimization for finding the first character of a
3685 non-anchored pattern, and for finding a character that is required later in the
3686 match were failing in some cases. This didn't break the matching; it just
3687 failed to optimize when it could. The way this is done has been re-implemented.
3688
3689 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3690
3691 28. Added a new feature that provides some of the functionality that Perl
3692 provides with (?{...}). The facility is termed a "callout". The way it is done
3693 in PCRE is for the caller to provide an optional function, by setting
3694 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3695 global variable. By default it is unset, which disables all calling out. To get
3696 the function called, the regex must include (?C) at appropriate points. This
3697 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3698 This provides a means of identifying different callout points. When PCRE
3699 reaches such a point in the regex, if pcre_callout has been set, the external
3700 function is called. It is provided with data in a structure called
3701 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3702 matching continues; if it returns a non-zero value, the match at the current
3703 point fails. However, backtracking will occur if possible. [This was changed
3704 later and other features added - see item 49 below.]
3705
3706 29. pcretest is upgraded to test the callout functionality. It provides a
3707 callout function that displays information. By default, it shows the start of
3708 the match and the current position in the text. There are some new data escapes
3709 to vary what happens:
3710
3711 \C+ in addition, show current contents of captured substrings
3712 \C- do not supply a callout function
3713 \C!n return 1 when callout number n is reached
3714 \C!n!m return 1 when callout number n is reached for the mth time
3715
3716 30. If pcregrep was called with the -l option and just a single file name, it
3717 output "<stdin>" if a match was found, instead of the file name.
3718
3719 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3720 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3721 pcre_exec(). This saves a malloc/free per call. The default value of
3722 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3723 when configuring.
3724
3725 32. The default maximum size of a compiled pattern is 64K. There have been a
3726 few cases of people hitting this limit. The code now uses macros to handle the
3727 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3728 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3729 configuring. Tests 2 and 5 work only with 2-byte links because they output
3730 debugging information about compiled patterns.
3731
3732 33. Internal code re-arrangements:
3733
3734 (a) Moved the debugging function for printing out a compiled regex into
3735 its own source file (printint.c) and used #include to pull it into
3736 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3737 separate copies.
3738
3739 (b) Defined the list of op-code names for debugging as a macro in
3740 internal.h so that it is next to the definition of the opcodes.
3741
3742 (c) Defined a table of op-code lengths for simpler skipping along compiled
3743 code. This is again a macro in internal.h so that it is next to the
3744 definition of the opcodes.
3745
3746 34. Added support for recursive calls to individual subpatterns, along the
3747 lines of Robin Houston's patch (but implemented somewhat differently).
3748
3749 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3750 allow it to read and process whole directories in Win32. This code was
3751 contributed by Lionel Fourquaux; it has not been tested by me.
3752
3753 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3754 used to name a group. Names consist of alphanumerics and underscores, and must
3755 be unique. Back references use the syntax (?P=name) and recursive calls use
3756 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3757 numbers. The function pcre_fullinfo() can be used after compilation to extract
3758 a name/number map. There are three relevant calls:
3759
3760 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3761 PCRE_INFO_NAMECOUNT yields the number of entries
3762 PCRE_INFO_NAMETABLE yields a pointer to the map.
3763
3764 The map is a vector of fixed-size entries. The size of each entry depends on
3765 the length of the longest name used. The first two bytes of each entry are the
3766 group number, most significant byte first. There follows the corresponding
3767 name, zero terminated. The names are in alphabetical order.
3768
3769 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3770 case instead of 255. Making it the same both with and without UTF-8 support
3771 means that the same test output works with both.
3772
3773 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3774 calling malloc() with a zero argument.
3775
3776 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3777 optimization. I've improved things by keeping a bitmap of backreferences with
3778 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3779 fact referenced, the optimization can be applied. It is unlikely that a
3780 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3781 the match to follow \n) will appear inside brackets with a number greater than
3782 31, but if it does, any back reference > 31 suppresses the optimization.
3783
3784 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3785 of disabling numbered capturing parentheses. Any opening parenthesis that is
3786 not followed by ? behaves as if it were followed by ?: but named parentheses
3787 can still be used for capturing (and they will acquire numbers in the usual
3788 way).
3789
3790 41. Redesigned the return codes from the match() function into yes/no/error so
3791 that errors can be passed back from deep inside the nested calls. A malloc
3792 failure while inside a recursive subpattern call now causes the
3793 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3794
3795 42. It is now possible to set a limit on the number of times the match()
3796 function is called in a call to pcre_exec(). This facility makes it possible to
3797 limit the amount of recursion and backtracking, though not in a directly
3798 obvious way, because the match() function is used in a number of different
3799 circumstances. The count starts from zero for each position in the subject
3800 string (for non-anchored patterns). The default limit is, for compatibility, a
3801 large number, namely 10 000 000. You can change this in two ways:
3802
3803 (a) When configuring PCRE before making, you can use --with-match-limit=n
3804 to set a default value for the compiled library.
3805
3806 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3807 a different value is set. See 45 below.
3808
3809 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3810
3811 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3812 of things that can be changed at compile time. The first argument specifies
3813 what is wanted and the second points to where the information is to be placed.
3814 The current list of available information is:
3815
3816 PCRE_CONFIG_UTF8
3817
3818 The output is an integer that is set to one if UTF-8 support is available;
3819 otherwise it is set to zero.
3820
3821 PCRE_CONFIG_NEWLINE
3822
3823 The output is an integer that it set to the value of the code that is used for
3824 newline. It is either LF (10) or CR (13).
3825
3826 PCRE_CONFIG_LINK_SIZE
3827
3828 The output is an integer that contains the number of bytes used for internal
3829 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3830
3831 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3832
3833 The output is an integer that contains the threshold above which the POSIX
3834 interface uses malloc() for output vectors. See item 31 above.
3835
3836 PCRE_CONFIG_MATCH_LIMIT
3837
3838 The output is an unsigned integer that contains the default limit of the number
3839 of match() calls in a pcre_exec() execution. See 42 above.
3840
3841 44. pcretest has been upgraded by the addition of the -C option. This causes it
3842 to extract all the available output from the new pcre_config() function, and to
3843 output it. The program then exits immediately.
3844
3845 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3846 order to support additional features. One way would have been to define
3847 pcre_exec2() (for example) with extra arguments, but this would not have been
3848 extensible, and would also have required all calls to the original function to
3849 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3850 is used for passing in "extra" data from pcre_study().
3851
3852 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3853 contains the following fields:
3854
3855 flags a bitmap indicating which of the following fields are set
3856 study_data opaque data from pcre_study()
3857 match_limit a way of specifying a limit on match() calls for a specific
3858 call to pcre_exec()
3859 callout_data data for callouts (see 49 below)
3860
3861 The flag bits are also defined in pcre.h, and are
3862
3863 PCRE_EXTRA_STUDY_DATA
3864 PCRE_EXTRA_MATCH_LIMIT
3865 PCRE_EXTRA_CALLOUT_DATA
3866
3867 The pcre_study() function now returns one of these new pcre_extra blocks, with
3868 the actual study data pointed to by the study_data field, and the
3869 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3870 before. That is, this change is entirely upwards-compatible and requires no
3871 change to existing code.
3872
3873 If you want to pass in additional data to pcre_exec(), you can either place it
3874 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3875 block.
3876
3877 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3878 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3879 times with different match limits, until it finds the minimum value needed for
3880 pcre_exec() to complete. The value is then output. This can be instructive; for
3881 most simple matches the number is quite small, but for pathological cases it
3882 gets very large very quickly.
3883
3884 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3885 returns the size of the data block pointed to by the study_data field in a
3886 pcre_extra block, that is, the value that was passed as the argument to
3887 pcre_malloc() when PCRE was getting memory in which to place the information
3888 created by pcre_study(). The fourth argument should point to a size_t variable.
3889 pcretest has been extended so that this information is shown after a successful
3890 pcre_study() call when information about the compiled regex is being displayed.
3891
3892 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3893 because what follows is always an absolute path. (Later: it turns out that this
3894 is more than cosmetic for MinGW, because it doesn't like empty path
3895 components.)
3896
3897 49. Some changes have been made to the callout feature (see 28 above):
3898
3899 (i) A callout function now has three choices for what it returns:
3900
3901 0 => success, carry on matching
3902 > 0 => failure at this point, but backtrack if possible
3903 < 0 => serious error, return this value from pcre_exec()
3904
3905 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3906 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3907 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3908 use by callout functions. It will never be used by PCRE itself.
3909
3910 (ii) The pcre_extra structure (see 45 above) has a void * field called
3911 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3912 pcre_callout_block structure has a field of the same name. The contents of
3913 the field passed in the pcre_extra structure are passed to the callout
3914 function in the corresponding field in the callout block. This makes it
3915 easier to use the same callout-containing regex from multiple threads. For
3916 testing, the pcretest program has a new data escape
3917
3918 \C*n pass the number n (may be negative) as callout_data
3919
3920 If the callout function in pcretest receives a non-zero value as
3921 callout_data, it returns that value.
3922
3923 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3924 there were some redundant $(CFLAGS) in commands that are now specified as
3925 $(LINK), which already includes $(CFLAGS).
3926
3927 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3928 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3929 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3930 one-byte characters throughout. Note that case-insensitive matching applies
3931 only to characters whose values are less than 256. PCRE doesn't support the
3932 notion of cases for higher-valued characters.
3933
3934 (i) A character class whose characters are all within 0-255 is handled as
3935 a bit map, and the map is inverted for negative classes. Previously, a
3936 character > 255 always failed to match such a class; however it should
3937 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3938
3939 (ii) A negated character class with a single character < 255 is coded as
3940 "not this character" (OP_NOT). This wasn't working properly when the test
3941 character was multibyte, either singly or repeated.
3942
3943 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3944 mode, for example: \x{100}{2,3}.
3945
3946 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3947 singly or repeated) now correctly test multibyte characters. However,
3948 PCRE doesn't recognize any characters with values greater than 255 as
3949 digits, spaces, or word characters. Such characters always match \D, \S,
3950 and \W, and never match \d, \s, or \w.
3951
3952 (v) Classes may now contain characters and character ranges with values
3953 greater than 255. For example: [ab\x{100}-\x{400}].
3954
3955 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3956 PCRE in UTF-8 mode.
3957
3958 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3959 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3960 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3961 value.)
3962
3963 53. The single man page has become too large. I have therefore split it up into
3964 a number of separate man pages. These also give rise to individual HTML pages;
3965 these are now put in a separate directory, and there is an index.html page that
3966 lists them all. Some hyperlinking between the pages has been installed.
3967
3968 54. Added convenience functions for handling named capturing parentheses.
3969
3970 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3971 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3972 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3973 are faulted.
3974
3975 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3976 calling configure. These values are used when compiling the dftables.c program
3977 which is run to generate the source of the default character tables. They
3978 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3979 you will need to set these values.
3980
3981 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3982
3983
3984 Version 3.9 02-Jan-02
3985 ---------------------
3986
3987 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3988
3989 2. If --disable-static was given, the building process failed when trying to
3990 build pcretest and pcregrep. (For some reason it was using libtool to compile
3991 them, which is not right, as they aren't part of the library.)
3992
3993
3994 Version 3.8 18-Dec-01
3995 ---------------------
3996
3997 1. The experimental UTF-8 code was completely screwed up. It was packing the
3998 bytes in the wrong order. How dumb can you get?
3999
4000
4001 Version 3.7 29-Oct-01
4002 ---------------------
4003
4004 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4005 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4006 this didn't happen under Solaris 8, where I normally test things.
4007
4008 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4009 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4010 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4011 this fix an hour or so after the initial 3.7 release.)
4012
4013
4014 Version 3.6 23-Oct-01
4015 ---------------------
4016
4017 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4018 offsets passed as NULL with zero offset count.
4019
4020 2. The config.guess and config.sub files had not been updated when I moved to
4021 the latest autoconf.
4022
4023
4024 Version 3.5 15-Aug-01
4025 ---------------------
4026
4027 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4028 had been forgotten.
4029
4030 2. By using declared but undefined structures, we can avoid using "void"
4031 definitions in pcre.h while keeping the internal definitions of the structures
4032 private.
4033
4034 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4035 user point of view, this means that both static and shared libraries are built
4036 by default, but this can be individually controlled. More of the work of
4037 handling this static/shared cases is now inside libtool instead of PCRE's make
4038 file.
4039
4040 4. The pcretest utility is now installed along with pcregrep because it is
4041 useful for users (to test regexs) and by doing this, it automatically gets
4042 relinked by libtool. The documentation has been turned into a man page, so
4043 there are now .1, .txt, and .html versions in /doc.
4044
4045 5. Upgrades to pcregrep:
4046 (i) Added long-form option names like gnu grep.
4047 (ii) Added --help to list all options with an explanatory phrase.
4048 (iii) Added -r, --recursive to recurse into sub-directories.
4049 (iv) Added -f, --file to read patterns from a file.
4050
4051 6. pcre_exec() was referring to its "code" argument before testing that
4052 argument for NULL (and giving an error if it was NULL).
4053
4054 7. Upgraded Makefile.in to allow for compiling in a different directory from
4055 the source directory.
4056
4057 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4058 options bits, the pointer it was passed was to an int instead of to an unsigned
4059 long int. This mattered only on 64-bit systems.
4060
4061 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4062 generated) instead of pcre.in, which it its source. Also made the same change
4063 in several of the .c files.
4064
4065 10. A new release of gcc defines printf() as a macro, which broke pcretest
4066 because it had an ifdef in the middle of a string argument for printf(). Fixed
4067 by using separate calls to printf().
4068
4069 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4070 script, to force use of CR or LF instead of \n in the source. On non-Unix
4071 systems, the value can be set in config.h.
4072
4073 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4074 absolute limit. Changed the text of the error message to make this clear, and
4075 likewise updated the man page.
4076
4077 13. The limit of 99 on the number of capturing subpatterns has been removed.
4078 The new limit is 65535, which I hope will not be a "real" limit.
4079
4080
4081 Version 3.4 22-Aug-00
4082 ---------------------
4083
4084 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4085
4086 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4087
4088
4089 Version 3.3 01-Aug-00
4090 ---------------------
4091
4092 1. If an octal character was given, but the value was greater than \377, it
4093 was not getting masked to the least significant bits, as documented. This could
4094 lead to crashes in some systems.
4095
4096 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4097 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4098
4099 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4100 These just pass their arguments on to (pcre_free)(), but they are provided
4101 because some uses of PCRE bind it to non-C systems that can call its functions,
4102 but cannot call free() or pcre_free() directly.
4103
4104 4. Add "make test" as a synonym for "make check". Corrected some comments in
4105 the Makefile.
4106
4107 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4108 Makefile.
4109
4110 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4111 command called pgrep for grepping around the active processes.
4112
4113 7. Added the beginnings of support for UTF-8 character strings.
4114
4115 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4116 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4117 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4118 out for the ar command.)
4119
4120
4121 Version 3.2 12-May-00
4122 ---------------------
4123
4124 This is purely a bug fixing release.
4125
4126 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4127 of ZA. This was just one example of several cases that could provoke this bug,
4128 which was introduced by change 9 of version 2.00. The code for breaking
4129 infinite loops after an iteration that matches an empty string was't working
4130 correctly.
4131
4132 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4133 when matched against abbab (for example). After matching an empty string, it
4134 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4135 caused it to match further down the string than it should.
4136
4137 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4138 was there because it doesn't seem to be needed, and it causes trouble on some
4139 systems, as it is not a Standard C header. It has been removed.
4140
4141 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4142 were reported on the Macintosh. The changes were from
4143
4144 while ((c = *(++ptr)) != 0 && c != '\n');
4145 to
4146 while ((c = *(++ptr)) != 0 && c != '\n') ;
4147
4148 Totally extraordinary, but if that's what it takes...
4149
4150 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4151 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4152 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4153 assumes the way PCRE uses memmove() (always moving upwards).
4154
4155 6. PCRE is being used in one environment where strchr() is not available. There
4156 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4157 faster code anyway.
4158
4159
4160 Version 3.1 09-Feb-00
4161 ---------------------
4162
4163 The only change in this release is the fixing of some bugs in Makefile.in for
4164 the "install" target:
4165
4166 (1) It was failing to install pcreposix.h.
4167
4168 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4169
4170
4171 Version 3.0 01-Feb-00
4172 ---------------------
4173
4174 1. Add support for the /+ modifier to perltest (to output $` like it does in
4175 pcretest).
4176
4177 2. Add support for the /g modifier to perltest.
4178
4179 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4180 matches null strings.
4181
4182 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4183 pattern. Perl treats empty patterns specially - it reuses the most recent
4184 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4185 effect.
4186
4187 5. The POSIX interface was broken in that it was just handing over the POSIX
4188 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4189 required a bigger vector, with some working space on the end. This means that
4190 the POSIX wrapper now has to get and free some memory, and copy the results.
4191
4192 6. Added some simple autoconf support, placing the test data and the
4193 documentation in separate directories, re-organizing some of the
4194 information files, and making it build pcre-config (a GNU standard). Also added
4195 libtool support for building PCRE as a shared library, which is now the
4196 default.
4197
4198 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4199 09 are not valid octal constants. Single digits will be used for minor values
4200 less than 10.
4201
4202 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4203 existing programs that set these in the POSIX interface can use PCRE without
4204 modification.
4205
4206 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4207 return all that pcre_info() returns, plus additional data. The pcre_info()
4208 function is retained for compatibility, but is considered to be obsolete.
4209
4210 10. Added experimental recursion feature (?R) to handle one common case that
4211 Perl 5.6 will be able to do with (?p{...}).
4212
4213 11. Added support for POSIX character classes like [:alpha:], which Perl is
4214 adopting.
4215
4216
4217 Version 2.08 31-Aug-99
4218 ----------------------
4219
4220 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4221 trying to match at the startoffset position, but instead was moving forward to
4222 the next newline as if a previous match had failed.
4223
4224 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4225 and could get into a loop if a null string was matched other than at the start
4226 of the subject.
4227
4228 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4229 be distinguished at compile time, and for completeness also added PCRE_DATE.
4230
4231 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4232 in GnuWin32 environments.
4233
4234
4235 Version 2.07 29-Jul-99
4236 ----------------------
4237
4238 1. The documentation is now supplied in plain text form and HTML as well as in
4239 the form of man page sources.
4240
4241 2. C++ compilers don't like assigning (void *) values to other pointer types.
4242 In particular this affects malloc(). Although there is no problem in Standard
4243 C, I've put in casts to keep C++ compilers happy.
4244
4245 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4246 should be (const char *).
4247
4248 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4249 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4250 However, I haven't made this a standard facility. The documentation doesn't
4251 mention it, and the Makefile doesn't support it.
4252
4253 5. The Makefile now contains an "install" target, with editable destinations at
4254 the top of the file. The pcretest program is not installed.
4255
4256 6. pgrep -V now gives the PCRE version number and date.
4257
4258 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4259 causing the entire string to be ignored, instead of just the last character.
4260
4261 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4262 non-matching string, it can take a very, very long time, even for strings of
4263 quite modest length, because of the nested recursion. PCRE now does better in
4264 some of these cases. It does this by remembering the last required literal
4265 character in the pattern, and pre-searching the subject to ensure it is present
4266 before running the real match. In other words, it applies a heuristic to detect
4267 some types of certain failure quickly, and in the above example, if presented
4268 with a string that has no trailing " it gives "no match" very quickly.
4269
4270 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4271 other alternatives are tried instead.
4272
4273
4274 Version 2.06 09-Jun-99
4275 ----------------------
4276
4277 1. Change pcretest's output for amount of store used to show just the code
4278 space, because the remainder (the data block) varies in size between 32-bit and
4279 64-bit systems.
4280
4281 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4282 start matching at. This allows lookbehinds to work when searching for multiple
4283 occurrences in a string.
4284
4285 3. Added additional options to pcretest for testing multiple occurrences:
4286
4287 /+ outputs the rest of the string that follows a match
4288 /g loops for multiple occurrences, using the new startoffset argument
4289 /G loops for multiple occurrences by passing an incremented pointer
4290
4291 4. PCRE wasn't doing the "first character" optimization for patterns starting
4292 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4293 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4294 the letter 'x'. On long subject strings, this gives a significant speed-up.
4295
4296
4297 Version 2.05 21-Apr-99
4298 ----------------------
4299
4300 1. Changed the type of magic_number from int to long int so that it works
4301 properly on 16-bit systems.
4302
4303 2. Fixed a bug which caused patterns starting with .* not to work correctly
4304 when the subject string contained newline characters. PCRE was assuming
4305 anchoring for such patterns in all cases, which is not correct because .* will
4306 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4307 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4308 must be retried after every newline in the subject.
4309
4310
4311 Version 2.04 18-Feb-99
4312 ----------------------
4313
4314 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4315 computation of the store needed to hold the pattern was incorrect (too large).
4316 If such patterns were nested a few deep, this could multiply and become a real
4317 problem.
4318
4319 2. Added /M option to pcretest to show the memory requirement of a specific
4320 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4321
4322 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4323 compiled in such a way that the backtracking after subsequent failure was
4324 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4325 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4326
4327
4328 Version 2.03 02-Feb-99
4329 ----------------------
4330
4331 1. Fixed typo and small mistake in man page.
4332
4333 2. Added 4th condition (GPL supersedes if conflict) and created separate
4334 LICENCE file containing the conditions.
4335
4336 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4337 Perl, that is the internal \ allows the delimiter to be included in the
4338 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4339 the final delimiter, add \ to the end of the pattern (to test the error).
4340
4341 4. Added the convenience functions for extracting substrings after a successful
4342 match. Updated pcretest to make it able to test these functions.
4343
4344
4345 Version 2.02 14-Jan-99
4346 ----------------------
4347
4348 1. Initialized the working variables associated with each extraction so that
4349 their saving and restoring doesn't refer to uninitialized store.
4350
4351 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4352 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4353 fix the problem.
4354
4355 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4356 calls, and wasn't printing the correct value for compiling calls. Increased the
4357 default value of LOOPREPEAT, and the number of significant figures in the
4358 times.
4359
4360 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4361
4362 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4363 a building problem on Windows NT with a FAT file system.
4364
4365
4366 Version 2.01 21-Oct-98
4367 ----------------------
4368
4369 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4370 to character tables built by pcre_maketables() in the current locale. If NULL
4371 is passed, the default tables are used.
4372
4373
4374 Version 2.00 24-Sep-98
4375 ----------------------
4376
4377 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4378 it any more.
4379
4380 2. Allow quantification of (?>) groups, and make it work correctly.
4381
4382 3. The first character computation wasn't working for (?>) groups.
4383
4384 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4385 end of the subject) and add 5.005's \z, which really does match only at the
4386 very end of the subject.
4387
4388 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4389
4390 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4391 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4392 localized options. All options to pcre_study() were also removed.
4393
4394 7. Add other new features from 5.005:
4395
4396 $(?<= positive lookbehind
4397 $(?<! negative lookbehind
4398 (?imsx-imsx) added the unsetting capability
4399 such a setting is global if at outer level; local otherwise
4400 (?imsx-imsx:) non-capturing groups with option setting
4401 (?(cond)re|re) conditional pattern matching
4402
4403 A backreference to itself in a repeated group matches the previous
4404 captured string.
4405
4406 8. General tidying up of studying (both automatic and via "study")
4407 consequential on the addition of new assertions.
4408
4409 9. As in 5.005, unlimited repeated groups that could match an empty substring
4410 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4411 runtime if any iteration does actually match an empty substring.
4412
4413 10. Include the RunTest script in the distribution.
4414
4415 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4416 discrepancies, some of which were old and were also with respect to 5.004. They
4417 have now been fixed.
4418
4419
4420 Version 1.09 28-Apr-98
4421 ----------------------
4422
4423 1. A negated single character class followed by a quantifier with a minimum
4424 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4425 program crashes, or just wrong answers. This did not apply to negated classes
4426 containing more than one character, or to minima other than one.
4427
4428
4429 Version 1.08 27-Mar-98
4430 ----------------------
4431
4432 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4433
4434 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4435 latter must appear before anything that relies on it in the pattern.
4436
4437
4438 Version 1.07 16-Feb-98
4439 ----------------------
4440
4441 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4442 repeat of a potentially empty string).
4443
4444
4445 Version 1.06 23-Jan-98
4446 ----------------------
4447
4448 1. Added Markus Oberhumer's little patches for C++.
4449
4450 2. Literal strings longer than 255 characters were broken.
4451
4452
4453 Version 1.05 23-Dec-97
4454 ----------------------
4455
4456 1. Negated character classes containing more than one character were failing if
4457 PCRE_CASELESS was set at run time.
4458
4459
4460 Version 1.04 19-Dec-97
4461 ----------------------
4462
4463 1. Corrected the man page, where some "const" qualifiers had been omitted.
4464
4465 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4466 input syntax.
4467
4468 3. Fixed memory leak which occurred when a regex with back references was
4469 matched with an offsets vector that wasn't big enough. The temporary memory
4470 that is used in this case wasn't being freed if the match failed.
4471
4472 4. Tidied pcretest to ensure it frees memory that it gets.
4473
4474 5. Temporary memory was being obtained in the case where the passed offsets
4475 vector was exactly big enough.
4476
4477 6. Corrected definition of offsetof() from change 5 below.
4478
4479 7. I had screwed up change 6 below and broken the rules for the use of
4480 setjmp(). Now fixed.
4481
4482
4483 Version 1.03 18-Dec-97
4484 ----------------------
4485
4486 1. A erroneous regex with a missing opening parenthesis was correctly
4487 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4488 on some systems.
4489
4490 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4491 it was reported that one broken compiler failed on the former because "code" is
4492 also an independent variable.
4493
4494 3. The erroneous regex a[]b caused an array overrun reference.
4495
4496 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4497 fail on data ending with that character. (It was going on too far, and checking
4498 the next character, typically a binary zero.) This was specific to the
4499 optimized code for single-character negative classes.
4500
4501 5. Added a contributed patch from the TIN world which does the following:
4502
4503 + Add an undef for memmove, in case the the system defines a macro for it.
4504
4505 + Add a definition of offsetof(), in case there isn't one. (I don't know
4506 the reason behind this - offsetof() is part of the ANSI standard - but
4507 it does no harm).
4508
4509 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4510 most of the places where whitespace preceded '#'. I have given up and
4511 allowed the remaining 2 cases to be at the margin.
4512
4513 + Rename some variables in pcre to eliminate shadowing. This seems very
4514 pedantic, but does no harm, of course.
4515
4516 6. Moved the call to setjmp() into its own function, to get rid of warnings
4517 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4518
4519 7. Constructs such as \d{8,} were compiling into the equivalent of
4520 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4521 outcome, but in this particular case used more store than had been allocated,
4522 which caused the bug to be discovered because it threw up an internal error.
4523
4524 8. The debugging code in both pcre and pcretest for outputting the compiled
4525 form of a regex was going wrong in the case of back references followed by
4526 curly-bracketed repeats.
4527
4528
4529 Version 1.02 12-Dec-97
4530 ----------------------
4531
4532 1. Typos in pcre.3 and comments in the source fixed.
4533
4534 2. Applied a contributed patch to get rid of places where it used to remove
4535 'const' from variables, and fixed some signed/unsigned and uninitialized
4536 variable warnings.
4537
4538 3. Added the "runtest" target to Makefile.
4539
4540 4. Set default compiler flag to -O2 rather than just -O.
4541
4542
4543 Version 1.01 19-Nov-97
4544 ----------------------
4545
4546 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4547 like /([ab]*)*/, that is, for classes with more than one character in them.
4548
4549 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4550 as /((?>a*))*/ (a PCRE_EXTRA facility).
4551
4552
4553 Version 1.00 18-Nov-97
4554 ----------------------
4555
4556 1. Added compile-time macros to support systems such as SunOS4 which don't have
4557 memmove() or strerror() but have other things that can be used instead.
4558
4559 2. Arranged that "make clean" removes the executables.
4560
4561
4562 Version 0.99 27-Oct-97
4563 ----------------------
4564
4565 1. Fixed bug in code for optimizing classes with only one character. It was
4566 initializing a 32-byte map regardless, which could cause it to run off the end
4567 of the memory it had got.
4568
4569 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4570
4571
4572 Version 0.98 22-Oct-97
4573 ----------------------
4574
4575 1. Fixed bug in code for handling temporary memory usage when there are more
4576 back references than supplied space in the ovector. This could cause segfaults.
4577
4578
4579 Version 0.97 21-Oct-97
4580 ----------------------
4581
4582 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4583
4584 2. Optimized negated single characters not to use a bit map.
4585
4586 3. Brought error texts together as macro definitions; clarified some of them;
4587 fixed one that was wrong - it said "range out of order" when it meant "invalid
4588 escape sequence".
4589
4590 4. Changed some char * arguments to const char *.
4591
4592 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4593
4594 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4595 pcretest.
4596
4597
4598 Version 0.96 16-Oct-97
4599 ----------------------
4600
4601 1. Added a simple "pgrep" utility to the distribution.
4602
4603 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4604 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4605 where "ddd" means "one or more decimal digits".
4606
4607 3. Fixed serious bug. If a pattern had a back reference, but the call to
4608 pcre_exec() didn't supply a large enough ovector to record the related
4609 identifying subpattern, the match always failed. PCRE now remembers the number
4610 of the largest back reference, and gets some temporary memory in which to save
4611 the offsets during matching if necessary, in order to ensure that
4612 backreferences always work.
4613
4614 4. Increased the compatibility with Perl in a number of ways:
4615
4616 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4617 to request this handling. The option can be set at compile or exec time.
4618
4619 (b) $ matches before a terminating newline by default; an option
4620 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4621 mode). The option can be set at compile or exec time.
4622
4623 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4624 the same as Perl's. If the decimal number it represents is less than 10
4625 or there aren't that many previous left capturing parentheses, an octal
4626 escape is read. Inside a character class, it's always an octal escape,
4627 even if it is a single digit.
4628
4629 (d) An escaped but undefined alphabetic character is taken as a literal,
4630 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4631 escapes.
4632
4633 (e) {0} is now permitted. (The previous item is removed from the compiled
4634 pattern).
4635
4636 5. Changed all the names of code files so that the basic parts are no longer
4637 than 10 characters, and abolished the teeny "globals.c" file.
4638
4639 6. Changed the handling of character classes; they are now done with a 32-byte
4640 bit map always.
4641
4642 7. Added the -d and /D options to pcretest to make it possible to look at the
4643 internals of compilation without having to recompile pcre.
4644
4645
4646 Version 0.95 23-Sep-97
4647 ----------------------
4648
4649 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4650 \x20 at the start of a run of normal characters. These were being treated as
4651 real characters, instead of the source characters being re-checked.
4652
4653
4654 Version 0.94 18-Sep-97
4655 ----------------------
4656
4657 1. The functions are now thread-safe, with the caveat that the global variables
4658 containing pointers to malloc() and free() or alternative functions are the
4659 same for all threads.
4660
4661 2. Get pcre_study() to generate a bitmap of initial characters for non-
4662 anchored patterns when this is possible, and use it if passed to pcre_exec().
4663
4664
4665 Version 0.93 15-Sep-97
4666 ----------------------
4667
4668 1. /(b)|(:+)/ was computing an incorrect first character.
4669
4670 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4671 but not actually doing anything yet.
4672
4673 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4674 as Perl does (e.g. [-az] or [az-]).
4675
4676 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4677 all possible positions.
4678
4679 5. Split up into different modules to avoid including unneeded functions in a
4680 compiled binary. However, compile and exec are still in one module. The "study"
4681 function is split off.
4682
4683 6. The character tables are now in a separate module whose source is generated
4684 by an auxiliary program - but can then be edited by hand if required. There are
4685 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4686 toupper() in the code.
4687
4688 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4689 make them global. Abolish the function for setting them, as the caller can now
4690 set them directly.
4691
4692
4693 Version 0.92 11-Sep-97
4694 ----------------------
4695
4696 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4697 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4698
4699 2. Caseless matching was not working in character classes if the characters in
4700 the pattern were in upper case.
4701
4702 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4703
4704 4. Make PCRE_ANCHORED public and accept as a compile option.
4705
4706 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4707 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4708 pass them.
4709
4710 6. Give an error if bad option bits passed at compile or run time.
4711
4712 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4713 pcretest to cause it to pass that flag.
4714
4715 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4716 options, and the first character, if set.
4717
4718 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4719
4720
4721 Version 0.91 10-Sep-97
4722 ----------------------
4723
4724 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4725 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4726
4727 2. PCRE was looping on encountering an indefinitely repeated back reference to
4728 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4729 Perl does - treats the match as successful.
4730
4731 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5