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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 958 - (show annotations) (download)
Wed Apr 11 10:19:10 2012 UTC (2 years, 8 months ago) by zherczeg
File size: 229092 byte(s)
support NO_START_OPTIMIZE in JIT
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.31
5 -----------------------------
6
7 1. Fixing a wrong JIT test case and some compiler warnings.
8
9 2. Removed a bashism from the RunTest script.
10
11 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
12 to unsigned type, result still unsigned" that was given by an MS compiler
13 on encountering the code "-sizeof(xxx)".
14
15 4. Partial matching support is added to the JIT compiler.
16
17 5. Fixed several bugs concerned with partial matching of items that consist
18 of more than one character:
19
20 (a) /^(..)\1/ did not partially match "aba" because checking references was
21 done on an "all or nothing" basis. This also applied to repeated
22 references.
23
24 (b) \R did not give a hard partial match if \r was found at the end of the
25 subject.
26
27 (c) \X did not give a hard partial match after matching one or more
28 characters at the end of the subject.
29
30 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
31 a partial match for the string "\r".
32
33 (e) When newline was set to CRLF, the metacharacter "." did not recognize
34 a partial match for a CR character at the end of the subject string.
35
36 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
37 running pcretest, the text "(JIT)" added to the output whenever JIT is
38 actually used to run the match.
39
40 7. Individual JIT compile options can be set in pcretest by following -s+[+]
41 or /S+[+] with a digit between 1 and 7.
42
43 8. OP_NOT now supports any UTF character not just single-byte ones.
44
45 9. (*MARK) control verb is now supported by the JIT compiler.
46
47 10. The command "./RunTest list" lists the available tests without actually
48 running any of them. (Because I keep forgetting what they all are.)
49
50 11. Add PCRE_INFO_MAXLOOKBEHIND.
51
52 12. Applied a (slightly modified) user-supplied patch that improves performance
53 when the heap is used for recursion (compiled with --disable-stack-for-
54 recursion). Instead of malloc and free for each heap frame each time a
55 logical recursion happens, frames are retained on a chain and re-used where
56 possible. This sometimes gives as much as 30% improvement.
57
58 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
59 call.
60
61 14. As documented, (*COMMIT) is now confined to within a positive assertion.
62
63 15. It is now possible to link pcretest with libedit as an alternative to
64 libreadline.
65
66 16. (*COMMIT) control verb is now supported by the JIT compiler.
67
68 17. The Unicode data tables have been updated to Unicode 6.1.0.
69
70 18. Added --file-list option to pcregrep.
71
72 19. Added binary file support to pcregrep, including the -a, --binary-files,
73 -I, and --text options.
74
75 20. The madvise function is renamed for posix_madvise for QNX compatibility
76 reasons. Fixed by Giuseppe D'Angelo.
77
78 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
79
80 22. Changed the option for creating symbolic links for 16-bit man pages from
81 -s to -sf so that re-installing does not cause issues.
82
83 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
84
85
86 Version 8.30 04-February-2012
87 -----------------------------
88
89 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
90 name is defined in ctype.h.
91
92 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
93 only in quite long subpatterns.
94
95 3. Removed the function pcre_info(), which has been obsolete and deprecated
96 since it was replaced by pcre_fullinfo() in February 2000.
97
98 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
99 match a (*MARK), and the match failed at the start of the subject, a
100 reference to memory before the start of the subject could occur. This bug
101 was introduced by fix 17 of release 8.21.
102
103 5. A reference to an unset group with zero minimum repetition was giving
104 totally wrong answers (in non-JavaScript-compatibility mode). For example,
105 /(another)?(\1?)test/ matched against "hello world test". This bug was
106 introduced in release 8.13.
107
108 6. Add support for 16-bit character strings (a large amount of work involving
109 many changes and refactorings).
110
111 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
112 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
113 from a file.
114
115 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
116 rounding is not applied in this particular case).
117
118 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
119 if they appear, or are escaped, in patterns.
120
121 10. Get rid of a number of -Wunused-but-set-variable warnings.
122
123 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
124 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
125 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
126 also returns the mark "x". This bug applied to capturing parentheses,
127 non-capturing parentheses, and atomic parentheses. It also applied to some
128 assertions.
129
130 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
131 information out of configure.ac instead of relying on pcre.h.generic, which
132 is not stored in the repository.
133
134 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
135 -lreadline.
136
137 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
138
139 15. Applied Graycode's patch to put the top-level frame on the stack rather
140 than the heap when not using the stack for recursion. This gives a
141 performance improvement in many cases when recursion is not deep.
142
143 16. Experimental code added to "pcretest -C" to output the stack frame size.
144
145
146 Version 8.21 12-Dec-2011
147 ------------------------
148
149 1. Updating the JIT compiler.
150
151 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
152 are added as well.
153
154 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
155 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
156 calling _pcre_jit_exec. Some extra comments are added.
157
158 4. (*MARK) settings inside atomic groups that do not contain any capturing
159 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
160 was introduced by change 18 for 8.20.
161
162 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
163 ECMA-262 standard.
164
165 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
166 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
167 This bug was probably introduced by change 9 of 8.13.
168
169 7. While fixing 6 above, I noticed that a number of other items were being
170 incorrectly rejected as "not fixed length". This arose partly because newer
171 opcodes had not been added to the fixed-length checking code. I have (a)
172 corrected the bug and added tests for these items, and (b) arranged for an
173 error to occur if an unknown opcode is encountered while checking for fixed
174 length instead of just assuming "not fixed length". The items that were
175 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
176 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
177 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
178
179 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
180 being incorrectly compiled and would have given unpredicatble results.
181
182 9. A possessively repeated subpattern with minimum repeat count greater than
183 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
184 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
185 the first (A) could occur when it should not.
186
187 10. Add a cast and remove a redundant test from the code.
188
189 11. JIT should use pcre_malloc/pcre_free for allocation.
190
191 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
192 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
193 is anything other than /usr, -L is still shown).
194
195 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
196
197 14. Perl does not support \N without a following name in a [] class; PCRE now
198 also gives an error.
199
200 15. If a forward reference was repeated with an upper limit of around 2000,
201 it caused the error "internal error: overran compiling workspace". The
202 maximum number of forward references (including repeats) was limited by the
203 internal workspace, and dependent on the LINK_SIZE. The code has been
204 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
205 the default depends on LINK_SIZE. There is a new upper limit (for safety)
206 of around 200,000 forward references. While doing this, I also speeded up
207 the filling in of repeated forward references.
208
209 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
210 incorrectly expecting the subject to contain another "a" after the start.
211
212 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
213 in the match, the SKIP should be ignored. This was not happening; instead
214 the SKIP was being treated as NOMATCH. For patterns such as
215 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
216 tested.
217
218 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
219 now much more compatible with Perl, in particular in cases where the result
220 is a non-match for a non-anchored pattern. For example, if
221 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
222 "m", where previously it did not return a name. A side effect of this
223 change is that for partial matches, the last encountered mark name is
224 returned, as for non matches. A number of tests that were previously not
225 Perl-compatible have been moved into the Perl-compatible test files. The
226 refactoring has had the pleasing side effect of removing one argument from
227 the match() function, thus reducing its stack requirements.
228
229 19. If the /S+ option was used in pcretest to study a pattern using JIT,
230 subsequent uses of /S (without +) incorrectly behaved like /S+.
231
232 21. Retrieve executable code size support for the JIT compiler and fixing
233 some warnings.
234
235 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
236 not work when the shorter character appeared right at the end of the
237 subject string.
238
239 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
240 systems.
241
242 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
243 output it when the /M option is used in pcretest.
244
245 25. The CheckMan script was not being included in the distribution. Also, added
246 an explicit "perl" to run Perl scripts from the PrepareRelease script
247 because this is reportedly needed in Windows.
248
249 26. If study data was being save in a file and studying had not found a set of
250 "starts with" bytes for the pattern, the data written to the file (though
251 never used) was taken from uninitialized memory and so caused valgrind to
252 complain.
253
254 27. Updated RunTest.bat as provided by Sheri Pierce.
255
256 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
257
258 29. Computation of memory usage for the table of capturing group names was
259 giving an unnecessarily large value.
260
261
262 Version 8.20 21-Oct-2011
263 ------------------------
264
265 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
266 a POSIX class. After further experiments with Perl, which convinced me that
267 Perl has bugs and confusions, a closing square bracket is no longer allowed
268 in a POSIX name. This bug also affected patterns with classes that started
269 with full stops.
270
271 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
272 captured substring, but while checking the failing first alternative,
273 substring 1 is temporarily captured. If the output vector supplied to
274 pcre_exec() was not big enough for this capture, the yield of the function
275 was still zero ("insufficient space for captured substrings"). This cannot
276 be totally fixed without adding another stack variable, which seems a lot
277 of expense for a edge case. However, I have improved the situation in cases
278 such as /(a)(b)x|abc/ matched against "abc", where the return code
279 indicates that fewer than the maximum number of slots in the ovector have
280 been set.
281
282 3. Related to (2) above: when there are more back references in a pattern than
283 slots in the output vector, pcre_exec() uses temporary memory during
284 matching, and copies in the captures as far as possible afterwards. It was
285 using the entire output vector, but this conflicts with the specification
286 that only 2/3 is used for passing back captured substrings. Now it uses
287 only the first 2/3, for compatibility. This is, of course, another edge
288 case.
289
290 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
291 main code base, and can be used by building with --enable-jit. When this is
292 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
293 runtime --no-jit option is given.
294
295 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
296 ovector, the return from the function was zero, implying that there were
297 other matches that did not fit. The correct "exactly full" value is now
298 returned.
299
300 6. If a subpattern that was called recursively or as a subroutine contained
301 (*PRUNE) or any other control that caused it to give a non-standard return,
302 invalid errors such as "Error -26 (nested recursion at the same subject
303 position)" or even infinite loops could occur.
304
305 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
306 computing the minimum length on reaching *ACCEPT, and so ended up with the
307 wrong value of 1 rather than 0. Further investigation indicates that
308 computing a minimum subject length in the presence of *ACCEPT is difficult
309 (think back references, subroutine calls), and so I have changed the code
310 so that no minimum is registered for a pattern that contains *ACCEPT.
311
312 8. If (*THEN) was present in the first (true) branch of a conditional group,
313 it was not handled as intended. [But see 16 below.]
314
315 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
316 Sheri Pierce.
317
318 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
319 the first byte in a match must be "a".
320
321 11. Change 17 for 8.13 increased the recursion depth for patterns like
322 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
323 pattern contains any instances of (*THEN). If it does not, the old
324 optimizations are restored. It would be nice to do this on a per-group
325 basis, but at the moment that is not feasible.
326
327 12. In some environments, the output of pcretest -C is CRLF terminated. This
328 broke RunTest's code that checks for the link size. A single white space
329 character after the value is now allowed for.
330
331 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
332 For "fr", it uses the Windows-specific input and output files.
333
334 14. If (*THEN) appeared in a group that was called recursively or as a
335 subroutine, it did not work as intended. [But see next item.]
336
337 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
338 pattern fragments (but not containing any | characters). If A and B are
339 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
340 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
341 D. In other words, Perl considers parentheses that do not contain any |
342 characters to be part of a surrounding alternative, whereas PCRE was
343 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
344 differently. PCRE now behaves in the same way as Perl, except in the case
345 of subroutine/recursion calls such as (?1) which have in any case always
346 been different (but PCRE had them first :-).
347
348 16. Related to 15 above: Perl does not treat the | in a conditional group as
349 creating alternatives. Such a group is treated in the same way as an
350 ordinary group without any | characters when processing (*THEN). PCRE has
351 been changed to match Perl's behaviour.
352
353 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
354 RunGrepTest script failed.
355
356 18. Change 22 for version 13 caused atomic groups to use more stack. This is
357 inevitable for groups that contain captures, but it can lead to a lot of
358 stack use in large patterns. The old behaviour has been restored for atomic
359 groups that do not contain any capturing parentheses.
360
361 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
362 suppress the check for a minimum subject length at run time. (If it was
363 given to pcre_exec() or pcre_dfa_exec() it did work.)
364
365 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
366 fail to work when decoding hex characters in data strings in EBCDIC
367 environments.
368
369 21. It appears that in at least one Mac OS environment, the isxdigit() function
370 is implemented as a macro that evaluates to its argument more than once,
371 contravening the C 90 Standard (I haven't checked a later standard). There
372 was an instance in pcretest which caused it to go wrong when processing
373 \x{...} escapes in subject strings. The has been rewritten to avoid using
374 things like p++ in the argument of isxdigit().
375
376
377 Version 8.13 16-Aug-2011
378 ------------------------
379
380 1. The Unicode data tables have been updated to Unicode 6.0.0.
381
382 2. Two minor typos in pcre_internal.h have been fixed.
383
384 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
385 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
386 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
387
388 4. There were a number of related bugs in the code for matching backrefences
389 caselessly in UTF-8 mode when codes for the characters concerned were
390 different numbers of bytes. For example, U+023A and U+2C65 are an upper
391 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
392 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
393 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
394 2-byte code at the end of the subject (it thought there wasn't enough data
395 left).
396
397 5. Comprehensive information about what went wrong is now returned by
398 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
399 as the output vector has at least 2 elements. The offset of the start of
400 the failing character and a reason code are placed in the vector.
401
402 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
403 now returned is for the first byte of the failing character, instead of the
404 last byte inspected. This is an incompatible change, but I hope it is small
405 enough not to be a problem. It makes the returned offset consistent with
406 pcre_exec() and pcre_dfa_exec().
407
408 7. pcretest now gives a text phrase as well as the error number when
409 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
410 failure, the offset and reason code are output.
411
412 8. When \R was used with a maximizing quantifier it failed to skip backwards
413 over a \r\n pair if the subsequent match failed. Instead, it just skipped
414 back over a single character (\n). This seems wrong (because it treated the
415 two characters as a single entity when going forwards), conflicts with the
416 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
417 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
418 has been changed.
419
420 9. Some internal refactoring has changed the processing so that the handling
421 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
422 time (the PCRE_DOTALL option was changed this way some time ago: version
423 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
424 which was always a bit of a fudge. It also means that there is one less
425 argument for the match() function, which reduces its stack requirements
426 slightly. This change also fixes an incompatibility with Perl: the pattern
427 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
428
429 10. More internal refactoring has drastically reduced the number of recursive
430 calls to match() for possessively repeated groups such as (abc)++ when
431 using pcre_exec().
432
433 11. While implementing 10, a number of bugs in the handling of groups were
434 discovered and fixed:
435
436 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
437 (a|)*(?1) gave a compile-time internal error.
438 ((a|)+)+ did not notice that the outer group could match an empty string.
439 (^a|^)+ was not marked as anchored.
440 (.*a|.*)+ was not marked as matching at start or after a newline.
441
442 12. Yet more internal refactoring has removed another argument from the match()
443 function. Special calls to this function are now indicated by setting a
444 value in a variable in the "match data" data block.
445
446 13. Be more explicit in pcre_study() instead of relying on "default" for
447 opcodes that mean there is no starting character; this means that when new
448 ones are added and accidentally left out of pcre_study(), testing should
449 pick them up.
450
451 14. The -s option of pcretest has been documented for ages as being an old
452 synonym of -m (show memory usage). I have changed it to mean "force study
453 for every regex", that is, assume /S for every regex. This is similar to -i
454 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
455 using it. It makes it easier to run collections of tests with and without
456 study enabled, and thereby test pcre_study() more easily. All the standard
457 tests are now run with and without -s (but some patterns can be marked as
458 "never study" - see 20 below).
459
460 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
461 restoration of the capturing data to the outer values was not happening
462 correctly.
463
464 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
465 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
466 pattern had matched an empty string, and so incorrectly returned a no
467 match.
468
469 17. There was optimizing code for the last branch of non-capturing parentheses,
470 and also for the obeyed branch of a conditional subexpression, which used
471 tail recursion to cut down on stack usage. Unfortunately, now that there is
472 the possibility of (*THEN) occurring in these branches, tail recursion is
473 no longer possible because the return has to be checked for (*THEN). These
474 two optimizations have therefore been removed. [But see 8.20/11 above.]
475
476 18. If a pattern containing \R was studied, it was assumed that \R always
477 matched two bytes, thus causing the minimum subject length to be
478 incorrectly computed because \R can also match just one byte.
479
480 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
481 was incorrectly computed.
482
483 20. If /S is present twice on a test pattern in pcretest input, it now
484 *disables* studying, thereby overriding the use of -s on the command line
485 (see 14 above). This is necessary for one or two tests to keep the output
486 identical in both cases.
487
488 21. When (*ACCEPT) was used in an assertion that matched an empty string and
489 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
490
491 22. When an atomic group that contained a capturing parenthesis was
492 successfully matched, but the branch in which it appeared failed, the
493 capturing was not being forgotten if a higher numbered group was later
494 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
495 group 1 to "a", when in fact it should be unset. This applied to multi-
496 branched capturing and non-capturing groups, repeated or not, and also to
497 positive assertions (capturing in negative assertions does not happen
498 in PCRE) and also to nested atomic groups.
499
500 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
501 subject after a captured substring, to make it easier to tell which of a
502 number of identical substrings has been captured.
503
504 24. The way atomic groups are processed by pcre_exec() has been changed so that
505 if they are repeated, backtracking one repetition now resets captured
506 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
507 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
508 "aaa". Previously, it would have been "a". As part of this code
509 refactoring, the way recursive calls are handled has also been changed.
510
511 25. If an assertion condition captured any substrings, they were not passed
512 back unless some other capturing happened later. For example, if
513 (?(?=(a))a) was matched against "a", no capturing was returned.
514
515 26. When studying a pattern that contained subroutine calls or assertions,
516 the code for finding the minimum length of a possible match was handling
517 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
518 group 1 called group 2 while simultaneously a separate group 2 called group
519 1). A stack overflow occurred in this case. I have fixed this by limiting
520 the recursion depth to 10.
521
522 27. Updated RunTest.bat in the distribution to the version supplied by Tom
523 Fortmann. This supports explicit test numbers on the command line, and has
524 argument validation and error reporting.
525
526 28. An instance of \X with an unlimited repeat could fail if at any point the
527 first character it looked at was a mark character.
528
529 29. Some minor code refactoring concerning Unicode properties and scripts
530 should reduce the stack requirement of match() slightly.
531
532 30. Added the '=' option to pcretest to check the setting of unused capturing
533 slots at the end of the pattern, which are documented as being -1, but are
534 not included in the return count.
535
536 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
537 compiled something random. Now it gives a compile-time error (as does
538 Perl).
539
540 32. A *MARK encountered during the processing of a positive assertion is now
541 recorded and passed back (compatible with Perl).
542
543 33. If --only-matching or --colour was set on a pcregrep call whose pattern
544 had alternative anchored branches, the search for a second match in a line
545 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
546 matched the line "0102" twice. The same bug affected patterns that started
547 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
548 twice.
549
550 34. Previously, PCRE did not allow quantification of assertions. However, Perl
551 does, and because of capturing effects, quantifying parenthesized
552 assertions may at times be useful. Quantifiers are now allowed for
553 parenthesized assertions.
554
555 35. A minor code tidy in pcre_compile() when checking options for \R usage.
556
557 36. \g was being checked for fancy things in a character class, when it should
558 just be a literal "g".
559
560 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
561 appearance of a nested POSIX class supersedes an apparent external class.
562 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
563 unescaped square brackets may also appear as part of class names. For
564 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
565 more like Perl. (But see 8.20/1 above.)
566
567 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
568 was because it thought it was \N{name}, which is not supported).
569
570 39. Add minix to OS list not supporting the -S option in pcretest.
571
572 40. PCRE tries to detect cases of infinite recursion at compile time, but it
573 cannot analyze patterns in sufficient detail to catch mutual recursions
574 such as ((?1))((?2)). There is now a runtime test that gives an error if a
575 subgroup is called recursively as a subpattern for a second time at the
576 same position in the subject string. In previous releases this might have
577 been caught by the recursion limit, or it might have run out of stack.
578
579 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
580 happen only once. PCRE was, however incorrectly giving a compile time error
581 "recursive call could loop indefinitely" because it cannot analyze the
582 pattern in sufficient detail. The compile time test no longer happens when
583 PCRE is compiling a conditional subpattern, but actual runaway loops are
584 now caught at runtime (see 40 above).
585
586 42. It seems that Perl allows any characters other than a closing parenthesis
587 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
588 has been changed to be the same.
589
590 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
591 as not to get warnings when autogen.sh is called. Also changed
592 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
593
594 44. To help people who use pcregrep to scan files containing exceedingly long
595 lines, the following changes have been made:
596
597 (a) The default value of the buffer size parameter has been increased from
598 8K to 20K. (The actual buffer used is three times this size.)
599
600 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
601 PCRE is built.
602
603 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
604 to be set at run time.
605
606 (d) Numerical values in pcregrep options can be followed by K or M, for
607 example --buffer-size=50K.
608
609 (e) If a line being scanned overflows pcregrep's buffer, an error is now
610 given and the return code is set to 2.
611
612 45. Add a pointer to the latest mark to the callout data block.
613
614 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
615 partial match of an empty string instead of no match. This was specific to
616 the use of ".".
617
618 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
619 complete match instead of a partial match. This bug was dependent on both
620 the PCRE_UTF8 and PCRE_DOTALL options being set.
621
622 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
623 starting byte set, because \b was not being ignored.
624
625
626 Version 8.12 15-Jan-2011
627 ------------------------
628
629 1. Fixed some typos in the markup of the man pages, and wrote a script that
630 checks for such things as part of the documentation building process.
631
632 2. On a big-endian 64-bit system, pcregrep did not correctly process the
633 --match-limit and --recursion-limit options (added for 8.11). In
634 particular, this made one of the standard tests fail. (The integer value
635 went into the wrong half of a long int.)
636
637 3. If the --colour option was given to pcregrep with -v (invert match), it
638 did strange things, either producing crazy output, or crashing. It should,
639 of course, ignore a request for colour when reporting lines that do not
640 match.
641
642 4. Another pcregrep bug caused similar problems if --colour was specified with
643 -M (multiline) and the pattern match finished with a line ending.
644
645 5. In pcregrep, when a pattern that ended with a literal newline sequence was
646 matched in multiline mode, the following line was shown as part of the
647 match. This seems wrong, so I have changed it.
648
649 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
650 the check for further matches in the same line (so they could be coloured)
651 to overrun the end of the current line. If another match was found, it was
652 incorrectly shown (and then shown again when found in the next line).
653
654 7. If pcregrep was compiled under Windows, there was a reference to the
655 function pcregrep_exit() before it was defined. I am assuming this was
656 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
657 reported by a user. I've moved the definition above the reference.
658
659
660 Version 8.11 10-Dec-2010
661 ------------------------
662
663 1. (*THEN) was not working properly if there were untried alternatives prior
664 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
665 backtracked to try for "b" instead of moving to the next alternative branch
666 at the same level (in this case, to look for "c"). The Perl documentation
667 is clear that when (*THEN) is backtracked onto, it goes to the "next
668 alternative in the innermost enclosing group".
669
670 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
671 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
672 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
673 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
674 (*THEN).
675
676 3. If \s appeared in a character class, it removed the VT character from
677 the class, even if it had been included by some previous item, for example
678 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
679 of \s, but is part of the POSIX "space" class.)
680
681 4. A partial match never returns an empty string (because you can always
682 match an empty string at the end of the subject); however the checking for
683 an empty string was starting at the "start of match" point. This has been
684 changed to the "earliest inspected character" point, because the returned
685 data for a partial match starts at this character. This means that, for
686 example, /(?<=abc)def/ gives a partial match for the subject "abc"
687 (previously it gave "no match").
688
689 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
690 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
691 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
692 has an implication that the given string is incomplete (because a partial
693 match is preferred over a full match). For this reason, these items now
694 give a partial match in this situation. [Aside: previously, the one case
695 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
696 match rather than a full match, which was wrong by the old rules, but is
697 now correct.]
698
699 6. There was a bug in the handling of #-introduced comments, recognized when
700 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
701 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
702 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
703 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
704 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
705 places in pcre_compile().
706
707 7. Related to (6) above, when pcre_compile() was skipping #-introduced
708 comments when looking ahead for named forward references to subpatterns,
709 the only newline sequence it recognized was NL. It now handles newlines
710 according to the set newline convention.
711
712 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
713 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
714 cater for a lack of strerror(). These oversights have been fixed.
715
716 9. Added --match-limit and --recursion-limit to pcregrep.
717
718 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
719
720 11. When the -o option was used, pcregrep was setting a return code of 1, even
721 when matches were found, and --line-buffered was not being honoured.
722
723 12. Added an optional parentheses number to the -o and --only-matching options
724 of pcregrep.
725
726 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
727 can match an empty string. The code to do it in pcretest and pcredemo
728 needed fixing:
729
730 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
731 only one byte after an empty string match just before CRLF (this case
732 just got forgotten; "any" and "anycrlf" were OK).
733
734 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
735 mode when an empty string match preceded an ASCII character followed by
736 a non-ASCII character. (The code for advancing by one character rather
737 than one byte was nonsense.)
738
739 (c) The pcredemo.c sample program did not have any code at all to handle
740 the cases when CRLF is a valid newline sequence.
741
742 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
743 as a starting offset was within the subject string. There is now a new
744 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
745 negative or greater than the length of the string. In order to test this,
746 pcretest is extended to allow the setting of negative starting offsets.
747
748 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
749 starting offset points to the beginning of a UTF-8 character was
750 unnecessarily clumsy. I tidied it up.
751
752 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
753 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
754
755 17. Nobody had reported that the --include_dir option, which was added in
756 release 7.7 should have been called --include-dir (hyphen, not underscore)
757 for compatibility with GNU grep. I have changed it to --include-dir, but
758 left --include_dir as an undocumented synonym, and the same for
759 --exclude-dir, though that is not available in GNU grep, at least as of
760 release 2.5.4.
761
762 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
763 characters from a string of bytes) have been redefined so as not to use
764 loops, in order to improve performance in some environments. At the same
765 time, I abstracted some of the common code into auxiliary macros to save
766 repetition (this should not affect the compiled code).
767
768 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
769 compile-time error is now given if \c is not followed by an ASCII
770 character, that is, a byte less than 128. (In EBCDIC mode, the code is
771 different, and any byte value is allowed.)
772
773 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
774 START_OPTIMIZE option, which is now allowed at compile time - but just
775 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
776 to pcregrep and other applications that have no direct access to PCRE
777 options. The new /Y option in pcretest sets this option when calling
778 pcre_compile().
779
780 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
781 back references. Groups containing recursive back references were forced to
782 be atomic by that change, but in the case of named groups, the amount of
783 memory required was incorrectly computed, leading to "Failed: internal
784 error: code overflow". This has been fixed.
785
786 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
787 pcretest.c, to avoid build problems in some Borland environments.
788
789
790 Version 8.10 25-Jun-2010
791 ------------------------
792
793 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
794 THEN.
795
796 2. (*ACCEPT) was not working when inside an atomic group.
797
798 3. Inside a character class, \B is treated as a literal by default, but
799 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
800 causes the error). The code is unchanged, but I tidied the documentation.
801
802 4. Inside a character class, PCRE always treated \R and \X as literals,
803 whereas Perl faults them if its -w option is set. I have changed PCRE so
804 that it faults them when PCRE_EXTRA is set.
805
806 5. Added support for \N, which always matches any character other than
807 newline. (It is the same as "." when PCRE_DOTALL is not set.)
808
809 6. When compiling pcregrep with newer versions of gcc which may have
810 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
811 declared with attribute warn_unused_result" were given. Just casting the
812 result to (void) does not stop the warnings; a more elaborate fudge is
813 needed. I've used a macro to implement this.
814
815 7. Minor change to pcretest.c to avoid a compiler warning.
816
817 8. Added four artifical Unicode properties to help with an option to make
818 \s etc use properties (see next item). The new properties are: Xan
819 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
820
821 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
822 use Unicode properties. (*UCP) at the start of a pattern can be used to set
823 this option. Modified pcretest to add /W to test this facility. Added
824 REG_UCP to make it available via the POSIX interface.
825
826 10. Added --line-buffered to pcregrep.
827
828 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
829 studied, and the match started with a letter with a code point greater than
830 127 whose first byte was different to the first byte of the other case of
831 the letter, the other case of this starting letter was not recognized
832 (#976).
833
834 12. If a pattern that was studied started with a repeated Unicode property
835 test, for example, \p{Nd}+, there was the theoretical possibility of
836 setting up an incorrect bitmap of starting bytes, but fortunately it could
837 not have actually happened in practice until change 8 above was made (it
838 added property types that matched character-matching opcodes).
839
840 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
841 possible starting bytes for non-anchored patterns.
842
843 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
844 \R, and also a number of cases that involve Unicode properties, both
845 explicit and implicit when PCRE_UCP is set.
846
847 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
848 input, it could crash or give wrong results if characters with values
849 greater than 0xc0 were present in the subject string. (Detail: it assumed
850 UTF-8 input when processing these items.)
851
852 16. Added a lot of (int) casts to avoid compiler warnings in systems where
853 size_t is 64-bit (#991).
854
855 17. Added a check for running out of memory when PCRE is compiled with
856 --disable-stack-for-recursion (#990).
857
858 18. If the last data line in a file for pcretest does not have a newline on
859 the end, a newline was missing in the output.
860
861 19. The default pcre_chartables.c file recognizes only ASCII characters (values
862 less than 128) in its various bitmaps. However, there is a facility for
863 generating tables according to the current locale when PCRE is compiled. It
864 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
865 characters, are recognized by isspace() and therefore were getting set in
866 these tables, and indeed these tables seem to approximate to ISO 8859. This
867 caused a problem in UTF-8 mode when pcre_study() was used to create a list
868 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
869 which of course cannot start UTF-8 characters. I have changed the code so
870 that only real ASCII characters (less than 128) and the correct starting
871 bytes for UTF-8 encodings are set for characters greater than 127 when in
872 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
873 altogether.)
874
875 20. Added the /T option to pcretest so as to be able to run tests with non-
876 standard character tables, thus making it possible to include the tests
877 used for 19 above in the standard set of tests.
878
879 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
880 reference to a subpattern the other side of a comment that contains an
881 opening parenthesis caused either an internal compiling error, or a
882 reference to the wrong subpattern.
883
884
885 Version 8.02 19-Mar-2010
886 ------------------------
887
888 1. The Unicode data tables have been updated to Unicode 5.2.0.
889
890 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
891 configured.
892
893 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
894 original author of that file, following a query about its status.
895
896 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
897 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
898
899 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
900 quantifier applied to a forward-referencing subroutine call, could compile
901 incorrect code or give the error "internal error: previously-checked
902 referenced subpattern not found".
903
904 6. Both MS Visual Studio and Symbian OS have problems with initializing
905 variables to point to external functions. For these systems, therefore,
906 pcre_malloc etc. are now initialized to local functions that call the
907 relevant global functions.
908
909 7. There were two entries missing in the vectors called coptable and poptable
910 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
911 I've fixed the data, and added a kludgy way of testing at compile time that
912 the lengths are correct (equal to the number of opcodes).
913
914 8. Following on from 7, I added a similar kludge to check the length of the
915 eint vector in pcreposix.c.
916
917 9. Error texts for pcre_compile() are held as one long string to avoid too
918 much relocation at load time. To find a text, the string is searched,
919 counting zeros. There was no check for running off the end of the string,
920 which could happen if a new error number was added without updating the
921 string.
922
923 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
924
925 11. \K was not working if it appeared in an atomic group or in a group that
926 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
927 \K is "not well defined" if used in an assertion. PCRE now accepts it if
928 the assertion is positive, but not if it is negative.
929
930 12. Change 11 fortuitously reduced the size of the stack frame used in the
931 "match()" function of pcre_exec.c by one pointer. Forthcoming
932 implementation of support for (*MARK) will need an extra pointer on the
933 stack; I have reserved it now, so that the stack frame size does not
934 decrease.
935
936 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
937 item in branch that calls a recursion is a subroutine call - as in the
938 second branch in the above example - was incorrectly given the compile-
939 time error "recursive call could loop indefinitely" because pcre_compile()
940 was not correctly checking the subroutine for matching a non-empty string.
941
942 14. The checks for overrunning compiling workspace could trigger after an
943 overrun had occurred. This is a "should never occur" error, but it can be
944 triggered by pathological patterns such as hundreds of nested parentheses.
945 The checks now trigger 100 bytes before the end of the workspace.
946
947 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
948
949
950 Version 8.01 19-Jan-2010
951 ------------------------
952
953 1. If a pattern contained a conditional subpattern with only one branch (in
954 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
955 computed the wrong minimum data length (which is of course zero for such
956 subpatterns). This could cause incorrect "no match" results.
957
958 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
959 the pattern is reset in the first branch, pcre_compile() failed with
960 "internal error: code overflow at offset...". This happened only when
961 the reset was to the original external option setting. (An optimization
962 abstracts leading options settings into an external setting, which was the
963 cause of this.)
964
965 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
966 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
967 assertion pattern did not match (meaning that the assertion was true), it
968 was incorrectly treated as false if the SKIP had been reached during the
969 matching. This also applied to assertions used as conditions.
970
971 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
972 assertion subpattern, including such a pattern used as a condition,
973 unpredictable results occurred, instead of the error return
974 PCRE_ERROR_DFA_UITEM.
975
976 5. The C++ GlobalReplace function was not working like Perl for the special
977 situation when an empty string is matched. It now does the fancy magic
978 stuff that is necessary.
979
980 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
981 removed. (These were left over from very, very early versions of PCRE.)
982
983 7. Some cosmetic changes to the code to make life easier when compiling it
984 as part of something else:
985
986 (a) Change DEBUG to PCRE_DEBUG.
987
988 (b) In pcre_compile(), rename the member of the "branch_chain" structure
989 called "current" as "current_branch", to prevent a collision with the
990 Linux macro when compiled as a kernel module.
991
992 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
993 prevent a collision with the Linux macro when compiled as a kernel
994 module.
995
996 8. In pcre_compile() there are some checks for integer overflows that used to
997 cast potentially large values to (double). This has been changed to that
998 when building, a check for int64_t is made, and if it is found, it is used
999 instead, thus avoiding the use of floating point arithmetic. (There is no
1000 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1001 double.
1002
1003 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1004 2005 (difference between two addresses compared to an unsigned value).
1005
1006 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1007 custom one, because of the following reported problem in Windows:
1008
1009 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1010 under Win32.
1011 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1012 therefore missing the function definition.
1013 - The compiler thus generates a "C" signature for the test function.
1014 - The linker fails to find the "C" function.
1015 - PCRE fails to configure if asked to do so against libbz2.
1016
1017 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1018 messages were output:
1019
1020 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1021 rerunning libtoolize, to keep the correct libtool macros in-tree.
1022 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1023
1024 I have done both of these things.
1025
1026 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1027 most of the time, it *can* run out if it is given a pattern that contains a
1028 runaway infinite recursion. I updated the discussion in the pcrestack man
1029 page.
1030
1031 13. Now that we have gone to the x.xx style of version numbers, the minor
1032 version may start with zero. Using 08 or 09 is a bad idea because users
1033 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1034 interpreted as invalid octal numbers. I've updated the previous comment in
1035 configure.ac, and also added a check that gives an error if 08 or 09 are
1036 used.
1037
1038 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1039 causing partial matching to fail when the end of the subject matched \W
1040 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1041
1042 15. There were some discrepancies between the declarations in pcre_internal.h
1043 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1044 their definitions. The declarations used "const uschar *" and the
1045 definitions used USPTR. Even though USPTR is normally defined as "const
1046 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1047 reported that: "This difference in casting confuses some C++ compilers, for
1048 example, SunCC recognizes above declarations as different functions and
1049 generates broken code for hbpcre." I have changed the declarations to use
1050 USPTR.
1051
1052 16. GNU libtool is named differently on some systems. The autogen.sh script now
1053 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1054 (FreeBSD).
1055
1056 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1057 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1058 comment: "Figure out how to create a longlong from a string: strtoll and
1059 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1060 instance, but it only takes 2 args instead of 3!"
1061
1062 18. A subtle bug concerned with back references has been fixed by a change of
1063 specification, with a corresponding code fix. A pattern such as
1064 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1065 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1066 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1067 same bug. Such groups have to be quantified to be useful, or contained
1068 inside another quantified group. (If there's no repetition, the reference
1069 can never match.) The problem arises because, having left the group and
1070 moved on to the rest of the pattern, a later failure that backtracks into
1071 the group uses the captured value from the final iteration of the group
1072 rather than the correct earlier one. I have fixed this in PCRE by forcing
1073 any group that contains a reference to itself to be an atomic group; that
1074 is, there cannot be any backtracking into it once it has completed. This is
1075 similar to recursive and subroutine calls.
1076
1077
1078 Version 8.00 19-Oct-09
1079 ----------------------
1080
1081 1. The table for translating pcre_compile() error codes into POSIX error codes
1082 was out-of-date, and there was no check on the pcre_compile() error code
1083 being within the table. This could lead to an OK return being given in
1084 error.
1085
1086 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1087 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1088 in a Windows environment.
1089
1090 3. The pcregrep --count option prints the count for each file even when it is
1091 zero, as does GNU grep. However, pcregrep was also printing all files when
1092 --files-with-matches was added. Now, when both options are given, it prints
1093 counts only for those files that have at least one match. (GNU grep just
1094 prints the file name in this circumstance, but including the count seems
1095 more useful - otherwise, why use --count?) Also ensured that the
1096 combination -clh just lists non-zero counts, with no names.
1097
1098 4. The long form of the pcregrep -F option was incorrectly implemented as
1099 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1100 but it seems right to fix it, and I didn't think it was worth preserving
1101 the old behaviour.
1102
1103 5. The command line items --regex=pattern and --regexp=pattern were not
1104 recognized by pcregrep, which required --regex pattern or --regexp pattern
1105 (with a space rather than an '='). The man page documented the '=' forms,
1106 which are compatible with GNU grep; these now work.
1107
1108 6. No libpcreposix.pc file was created for pkg-config; there was just
1109 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1110
1111 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1112 when UCP support is not needed, by modifying the Python script that
1113 generates it from Unicode data files. This should not matter if the module
1114 is correctly used as a library, but I received one complaint about 50K of
1115 unwanted data. My guess is that the person linked everything into his
1116 program rather than using a library. Anyway, it does no harm.
1117
1118 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1119 was a minimum greater than 1 for a wide character in a possessive
1120 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1121 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1122 character. Chaos in the form of incorrect output or a compiling loop could
1123 result.
1124
1125 9. The restrictions on what a pattern can contain when partial matching is
1126 requested for pcre_exec() have been removed. All patterns can now be
1127 partially matched by this function. In addition, if there are at least two
1128 slots in the offset vector, the offset of the earliest inspected character
1129 for the match and the offset of the end of the subject are set in them when
1130 PCRE_ERROR_PARTIAL is returned.
1131
1132 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1133 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1134 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1135 and may be more useful for multi-segment matching.
1136
1137 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1138 used to be given if ever the end of the subject was reached; now it is
1139 given only if matching could not proceed because another character was
1140 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1141 string "Z", which now yields "no match" instead of "partial match". In the
1142 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1143 final character ended with (*FAIL).
1144
1145 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1146 if the pattern had a "must contain" character that was already found in the
1147 earlier partial match, unless partial matching was again requested. For
1148 example, with the pattern /dog.(body)?/, the "must contain" character is
1149 "g". If the first part-match was for the string "dog", restarting with
1150 "sbody" failed. This bug has been fixed.
1151
1152 13. The string returned by pcre_dfa_exec() after a partial match has been
1153 changed so that it starts at the first inspected character rather than the
1154 first character of the match. This makes a difference only if the pattern
1155 starts with a lookbehind assertion or \b or \B (\K is not supported by
1156 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1157 matching functions compatible, and I think it's the right thing to do.
1158
1159 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1160 so that the demonstration program is easily available in environments where
1161 PCRE has not been installed from source.
1162
1163 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1164 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1165 library.
1166
1167 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1168 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1169 is not the first non-POSIX option to be added. Clearly some people find
1170 these options useful.
1171
1172 17. If a caller to the POSIX matching function regexec() passes a non-zero
1173 value for nmatch with a NULL value for pmatch, the value of
1174 nmatch is forced to zero.
1175
1176 18. RunGrepTest did not have a test for the availability of the -u option of
1177 the diff command, as RunTest does. It now checks in the same way as
1178 RunTest, and also checks for the -b option.
1179
1180 19. If an odd number of negated classes containing just a single character
1181 interposed, within parentheses, between a forward reference to a named
1182 subpattern and the definition of the subpattern, compilation crashed with
1183 an internal error, complaining that it could not find the referenced
1184 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1185 [The bug was that it was starting one character too far in when skipping
1186 over the character class, thus treating the ] as data rather than
1187 terminating the class. This meant it could skip too much.]
1188
1189 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1190 /g option in pcretest when the pattern contains \K, which makes it possible
1191 to have an empty string match not at the start, even when the pattern is
1192 anchored. Updated pcretest and pcredemo to use this option.
1193
1194 21. If the maximum number of capturing subpatterns in a recursion was greater
1195 than the maximum at the outer level, the higher number was returned, but
1196 with unset values at the outer level. The correct (outer level) value is
1197 now given.
1198
1199 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1200 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1201 make it do so. The string so far is captured, making this feature
1202 compatible with Perl.
1203
1204 23. The tests have been re-organized, adding tests 11 and 12, to make it
1205 possible to check the Perl 5.10 features against Perl 5.10.
1206
1207 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1208 pattern matches a fixed length string. PCRE did not allow this; now it
1209 does. Neither allows recursion.
1210
1211 25. I finally figured out how to implement a request to provide the minimum
1212 length of subject string that was needed in order to match a given pattern.
1213 (It was back references and recursion that I had previously got hung up
1214 on.) This code has now been added to pcre_study(); it finds a lower bound
1215 to the length of subject needed. It is not necessarily the greatest lower
1216 bound, but using it to avoid searching strings that are too short does give
1217 some useful speed-ups. The value is available to calling programs via
1218 pcre_fullinfo().
1219
1220 26. While implementing 25, I discovered to my embarrassment that pcretest had
1221 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1222 study optimizations had never been tested with that matching function.
1223 Oops. What is worse, even when it was passed study data, there was a bug in
1224 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1225 were also very few tests of studied patterns with pcre_dfa_exec().
1226
1227 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1228 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1229 on the other side of the coin, they are no longer allowed to have different
1230 names, because these cannot be distinguished in PCRE, and this has caused
1231 confusion. (This is a difference from Perl.)
1232
1233 28. When duplicate subpattern names are present (necessarily with different
1234 numbers, as required by 27 above), and a test is made by name in a
1235 conditional pattern, either for a subpattern having been matched, or for
1236 recursion in such a pattern, all the associated numbered subpatterns are
1237 tested, and the overall condition is true if the condition is true for any
1238 one of them. This is the way Perl works, and is also more like the way
1239 testing by number works.
1240
1241
1242 Version 7.9 11-Apr-09
1243 ---------------------
1244
1245 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1246 (pcretest), all targets were linked against these libraries. This included
1247 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1248 libraries. This caused unwanted dependencies to be created. This problem
1249 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1250 pcretest is linked with readline.
1251
1252 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1253 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1254 moved outside it again, because FALSE and TRUE are already defined in AIX,
1255 but BOOL is not.
1256
1257 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1258 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1259
1260 4. The pcregrep documentation said spaces were inserted as well as colons (or
1261 hyphens) following file names and line numbers when outputting matching
1262 lines. This is not true; no spaces are inserted. I have also clarified the
1263 wording for the --colour (or --color) option.
1264
1265 5. In pcregrep, when --colour was used with -o, the list of matching strings
1266 was not coloured; this is different to GNU grep, so I have changed it to be
1267 the same.
1268
1269 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1270 each matching line was coloured. Now it goes on to look for further matches
1271 of any of the test patterns, which is the same behaviour as GNU grep.
1272
1273 7. A pattern that could match an empty string could cause pcregrep to loop; it
1274 doesn't make sense to accept an empty string match in pcregrep, so I have
1275 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1276 seems to be how GNU grep behaves.
1277
1278 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1279 start or after a newline", because the conditional assertion was not being
1280 correctly handled. The rule now is that both the assertion and what follows
1281 in the first alternative must satisfy the test.
1282
1283 9. If auto-callout was enabled in a pattern with a conditional group whose
1284 condition was an assertion, PCRE could crash during matching, both with
1285 pcre_exec() and pcre_dfa_exec().
1286
1287 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1288 used for matching.
1289
1290 11. Unicode property support in character classes was not working for
1291 characters (bytes) greater than 127 when not in UTF-8 mode.
1292
1293 12. Added the -M command line option to pcretest.
1294
1295 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1296
1297 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1298
1299 16. Added comments and documentation about mis-use of no_arg in the C++
1300 wrapper.
1301
1302 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1303 from Martin Jerabek that uses macro names for all relevant character and
1304 string constants.
1305
1306 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1307 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1308 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1309 these, but not everybody uses configure.
1310
1311 19. A conditional group that had only one branch was not being correctly
1312 recognized as an item that could match an empty string. This meant that an
1313 enclosing group might also not be so recognized, causing infinite looping
1314 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1315 with the subject "ab", where knowledge that the repeated group can match
1316 nothing is needed in order to break the loop.
1317
1318 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1319 exec(), but without supplying a callout function, matching went wrong.
1320
1321 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1322 leak if the size of the offset vector was greater than 30. When the vector
1323 is smaller, the saved offsets during recursion go onto a local stack
1324 vector, but for larger vectors malloc() is used. It was failing to free
1325 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1326 error, in fact).
1327
1328 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1329 heapframe that is used only when UTF-8 support is enabled. This caused no
1330 problem, but was untidy.
1331
1332 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1333 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1334 included within another project.
1335
1336 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1337 slightly modified by me:
1338
1339 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1340 not building pcregrep.
1341
1342 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1343 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1344
1345 25. Forward references, both numeric and by name, in patterns that made use of
1346 duplicate group numbers, could behave incorrectly or give incorrect errors,
1347 because when scanning forward to find the reference group, PCRE was not
1348 taking into account the duplicate group numbers. A pattern such as
1349 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1350
1351 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1352 the feature of a custom pointer more persuasive (as requested by a user).
1353
1354 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1355 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1356 pre-defined.
1357
1358 28. Added support for (*UTF8) at the start of a pattern.
1359
1360 29. Arrange for flags added by the "release type" setting in CMake to be shown
1361 in the configuration summary.
1362
1363
1364 Version 7.8 05-Sep-08
1365 ---------------------
1366
1367 1. Replaced UCP searching code with optimized version as implemented for Ad
1368 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1369 stage table and inline lookup instead of a function, giving speed ups of 2
1370 to 5 times on some simple patterns that I tested. Permission was given to
1371 distribute the MultiStage2.py script that generates the tables (it's not in
1372 the tarball, but is in the Subversion repository).
1373
1374 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1375 scripts.
1376
1377 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1378 a group with a zero qualifier. The result of the study could be incorrect,
1379 or the function might crash, depending on the pattern.
1380
1381 4. Caseless matching was not working for non-ASCII characters in back
1382 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1383 It now works when Unicode Property Support is available.
1384
1385 5. In pcretest, an escape such as \x{de} in the data was always generating
1386 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1387 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1388 truncation.
1389
1390 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1391
1392 7. Added two (int) casts to pcregrep when printing the difference of two
1393 pointers, in case they are 64-bit values.
1394
1395 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1396 test 2 if it fails.
1397
1398 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1399 and a #define of that name to empty if it is not externally set. This is to
1400 allow users of MSVC to set it if necessary.
1401
1402 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1403 the convenience functions in the pcre_get.c source file.
1404
1405 11. An option change at the start of a pattern that had top-level alternatives
1406 could cause overwriting and/or a crash. This command provoked a crash in
1407 some environments:
1408
1409 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1410
1411 This potential security problem was recorded as CVE-2008-2371.
1412
1413 12. For a pattern where the match had to start at the beginning or immediately
1414 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1415 pcre_dfa_exec() could read past the end of the passed subject if there was
1416 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1417 pcretest so that it places the subject at the end of its malloc-ed buffer.
1418
1419 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1420 exec() might read past the end of the data buffer in UTF-8 mode.
1421
1422 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1423 the data contained the byte 0x85 as part of a UTF-8 character within its
1424 first line. This applied both to normal and DFA matching.
1425
1426 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1427 /^[^d]*?$/8 failed to match "abc".
1428
1429 16. Added a missing copyright notice to pcrecpp_internal.h.
1430
1431 17. Make it more clear in the documentation that values returned from
1432 pcre_exec() in ovector are byte offsets, not character counts.
1433
1434 18. Tidied a few places to stop certain compilers from issuing warnings.
1435
1436 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1437 supplied by Stefan Weber. I made a further small update for 7.8 because
1438 there is a change of source arrangements: the pcre_searchfuncs.c module is
1439 replaced by pcre_ucd.c.
1440
1441
1442 Version 7.7 07-May-08
1443 ---------------------
1444
1445 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1446 a string to a long long, pretend we don't even have a long long." This is
1447 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1448
1449 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1450 pre-7.6 versions, which defined a global no_arg variable instead of putting
1451 it in the RE class. (See also #8 below.)
1452
1453 3. Remove a line of dead code, identified by coverity and reported by Nuno
1454 Lopes.
1455
1456 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1457
1458 (1) The include/exclude patterns were being applied to the whole pathnames
1459 of files, instead of just to the final components.
1460
1461 (2) If there was more than one level of directory, the subdirectories were
1462 skipped unless they satisfied the include/exclude conditions. This is
1463 inconsistent with GNU grep (and could even be seen as contrary to the
1464 pcregrep specification - which I improved to make it absolutely clear).
1465 The action now is always to scan all levels of directory, and just
1466 apply the include/exclude patterns to regular files.
1467
1468 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1469 --exclude_dir in the tests to avoid scanning .svn directories.
1470
1471 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1472 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1473 doesn't support NULs in patterns.
1474
1475 7. Added some missing "const"s to declarations of static tables in
1476 pcre_compile.c and pcre_dfa_exec.c.
1477
1478 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1479 caused by fix #2 above. (Subsequently also a second patch to fix the
1480 first patch. And a third patch - this was a messy problem.)
1481
1482 9. Applied Craig's patch to remove the use of push_back().
1483
1484 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1485 matching function regexec().
1486
1487 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1488 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1489 references. PCRE supports relative numbers with this syntax (I don't think
1490 Oniguruma does).
1491
1492 12. Previously, a group with a zero repeat such as (...){0} was completely
1493 omitted from the compiled regex. However, this means that if the group
1494 was called as a subroutine from elsewhere in the pattern, things went wrong
1495 (an internal error was given). Such groups are now left in the compiled
1496 pattern, with a new opcode that causes them to be skipped at execution
1497 time.
1498
1499 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1500 to the way PCRE behaves:
1501
1502 (a) A lone ] character is dis-allowed (Perl treats it as data).
1503
1504 (b) A back reference to an unmatched subpattern matches an empty string
1505 (Perl fails the current match path).
1506
1507 (c) A data ] in a character class must be notated as \] because if the
1508 first data character in a class is ], it defines an empty class. (In
1509 Perl it is not possible to have an empty class.) The empty class []
1510 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1511 The negative empty class [^] matches any one character, independently
1512 of the DOTALL setting.
1513
1514 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1515 non-existent subpattern following a character class starting with ']' and
1516 containing () gave an internal compiling error instead of "reference to
1517 non-existent subpattern". Fortunately, when the pattern did exist, the
1518 compiled code was correct. (When scanning forwards to check for the
1519 existencd of the subpattern, it was treating the data ']' as terminating
1520 the class, so got the count wrong. When actually compiling, the reference
1521 was subsequently set up correctly.)
1522
1523 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1524 it was being rejected as not supported by pcre_dfa_exec(), even though
1525 other assertions are supported. I have made pcre_dfa_exec() support
1526 (*FAIL).
1527
1528 16. The implementation of 13c above involved the invention of a new opcode,
1529 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1530 cannot be changed at match time, I realized I could make a small
1531 improvement to matching performance by compiling OP_ALLANY instead of
1532 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1533 on the OP_ANY path.
1534
1535 17. Compiling pcretest on Windows with readline support failed without the
1536 following two fixes: (1) Make the unistd.h include conditional on
1537 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1538
1539 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1540 ncurses library to be included for pcretest when ReadLine support is
1541 requested, but also to allow for it to be overridden. This patch came from
1542 Daniel Bergström.
1543
1544 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1545 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1546 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1547 spotting this.
1548
1549
1550 Version 7.6 28-Jan-08
1551 ---------------------
1552
1553 1. A character class containing a very large number of characters with
1554 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1555 overflow.
1556
1557 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1558 HAVE_LONG_LONG is not defined.
1559
1560 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1561 bring them up to date and include new features. This patch includes:
1562
1563 - Fixed PH's badly added libz and libbz2 support.
1564 - Fixed a problem with static linking.
1565 - Added pcredemo. [But later removed - see 7 below.]
1566 - Fixed dftables problem and added an option.
1567 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1568 HAVE_LONG_LONG.
1569 - Added readline support for pcretest.
1570 - Added an listing of the option settings after cmake has run.
1571
1572 4. A user submitted a patch to Makefile that makes it easy to create
1573 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1574 Makefile.am that cause it to include this special target, without
1575 affecting anything else. Note that the same mingw target plus all
1576 the other distribution libraries and programs are now supported
1577 when configuring with CMake (see 6 below) instead of with
1578 Configure/Make.
1579
1580 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1581 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1582 exported in the Windows port". It has not yet been confirmed that the patch
1583 solves the problem, but it does no harm.
1584
1585 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1586 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1587 with CMake, and also correct the comment about stack recursion.
1588
1589 7. Remove the automatic building of pcredemo from the ./configure system and
1590 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1591 of a program that users should build themselves after PCRE is installed, so
1592 building it automatically is not really right. What is more, it gave
1593 trouble in some build environments.
1594
1595 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1596
1597
1598 Version 7.5 10-Jan-08
1599 ---------------------
1600
1601 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1602 values in parens when parsing an RE using the C++ wrapper."
1603
1604 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1605 Characters greater than 255 were excluded from the class instead of being
1606 included.
1607
1608 3. The same bug as (2) above applied to negated POSIX classes such as
1609 [:^space:].
1610
1611 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1612 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1613 I have changed it.
1614
1615 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1616 first named subpattern) and a construct such as (?&a) would reference the
1617 first named subpattern whose name started with "a" (in other words, the
1618 length check was missing). Both these problems are fixed. "Subpattern name
1619 expected" is now given for (?&) (a zero-length name), and this patch also
1620 makes it give the same error for \k'' (previously it complained that that
1621 was a reference to a non-existent subpattern).
1622
1623 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1624 this is right because (?- can be followed by option settings as well as by
1625 digits. I have, however, made the messages clearer.
1626
1627 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1628 than the number used in the conditional) now cause a compile-time error.
1629 This is actually not compatible with Perl, which accepts such patterns, but
1630 treats the conditional as always being FALSE (as PCRE used to), but it
1631 seems to me that giving a diagnostic is better.
1632
1633 8. Change "alphameric" to the more common word "alphanumeric" in comments
1634 and messages.
1635
1636 9. Fix two occurrences of "backslash" in comments that should have been
1637 "backspace".
1638
1639 10. Remove two redundant lines of code that can never be obeyed (their function
1640 was moved elsewhere).
1641
1642 11. The program that makes PCRE's Unicode character property table had a bug
1643 which caused it to generate incorrect table entries for sequences of
1644 characters that have the same character type, but are in different scripts.
1645 It amalgamated them into a single range, with the script of the first of
1646 them. In other words, some characters were in the wrong script. There were
1647 thirteen such cases, affecting characters in the following ranges:
1648
1649 U+002b0 - U+002c1
1650 U+0060c - U+0060d
1651 U+0061e - U+00612
1652 U+0064b - U+0065e
1653 U+0074d - U+0076d
1654 U+01800 - U+01805
1655 U+01d00 - U+01d77
1656 U+01d9b - U+01dbf
1657 U+0200b - U+0200f
1658 U+030fc - U+030fe
1659 U+03260 - U+0327f
1660 U+0fb46 - U+0fbb1
1661 U+10450 - U+1049d
1662
1663 12. The -o option (show only the matching part of a line) for pcregrep was not
1664 compatible with GNU grep in that, if there was more than one match in a
1665 line, it showed only the first of them. It now behaves in the same way as
1666 GNU grep.
1667
1668 13. If the -o and -v options were combined for pcregrep, it printed a blank
1669 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1670 does the same. The return code can be used to tell if there were any
1671 non-matching lines.
1672
1673 14. Added --file-offsets and --line-offsets to pcregrep.
1674
1675 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1676 infinitely looping recursion. The bug was that positive lookaheads were not
1677 being skipped when checking for a possible empty match (negative lookaheads
1678 and both kinds of lookbehind were skipped).
1679
1680 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1681 inclusion of <windows.h> to before rather than after the definition of
1682 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1683
1684 17. Specifying a possessive quantifier with a specific limit for a Unicode
1685 character property caused pcre_compile() to compile bad code, which led at
1686 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1687 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1688 caused the error; without that there was no problem.
1689
1690 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1691
1692 19. Added --enable-pcretest-libreadline.
1693
1694 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1695 RE::GlobalReplace(). As a result, the number of replacements returned was
1696 double what it should be. I removed one of the increments, but Craig sent a
1697 later patch that removed the other one (the right fix) and added unit tests
1698 that check the return values (which was not done before).
1699
1700 21. Several CMake things:
1701
1702 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1703 the names libpcre and libpcreposix, not just pcre and pcreposix.
1704
1705 (2) The above change means that pcretest and pcregrep are now correctly
1706 linked with the newly-built libraries, not previously installed ones.
1707
1708 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1709
1710 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1711 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1712 UTF-8 newline character). The key issue is that the pattern starts .*;
1713 this means that the match must be either at the beginning, or after a
1714 newline. The bug was in the code for advancing after a failed match and
1715 checking that the new position followed a newline. It was not taking
1716 account of UTF-8 characters correctly.
1717
1718 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1719 character classes. PCRE was not treating the sequence [:...:] as a
1720 character class unless the ... were all letters. Perl, however, seems to
1721 allow any characters between [: and :], though of course it rejects as
1722 unknown any "names" that contain non-letters, because all the known class
1723 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1724 for example, whereas PCRE did not - it did not recognize a POSIX character
1725 class. This seemed a bit dangerous, so the code has been changed to be
1726 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1727 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1728 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1729 Perl does, and where it didn't before.
1730
1731 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1732 Windows environments %n is disabled by default.
1733
1734
1735 Version 7.4 21-Sep-07
1736 ---------------------
1737
1738 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1739 means that a class such as [\s] counted as "explicit reference to CR or
1740 LF". That isn't really right - the whole point of the change was to try to
1741 help when there was an actual mention of one of the two characters. So now
1742 the change happens only if \r or \n (or a literal CR or LF) character is
1743 encountered.
1744
1745 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1746 of both had grown to the point where there were only 3 bits left.
1747 Fortunately, there was spare space in the data structure, and so I have
1748 moved the internal flags into a new 16-bit field to free up more option
1749 bits.
1750
1751 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1752 but did not set the internal JCHANGED flag - either of these is enough to
1753 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1754 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1755 start sets both bits.
1756
1757 4. Added options (at build time, compile time, exec time) to change \R from
1758 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1759
1760 5. doc/pcresyntax.html was missing from the distribution.
1761
1762 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1763 compatibility, even though it is no longer used.
1764
1765 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1766 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1767 windows.h file is present (where different names are used). [This was
1768 reversed later after testing - see 16 below.]
1769
1770 8. Changed all #include <config.h> to #include "config.h". There were also
1771 some further <pcre.h> cases that I changed to "pcre.h".
1772
1773 9. When pcregrep was used with the --colour option, it missed the line ending
1774 sequence off the lines that it output.
1775
1776 10. It was pointed out to me that arrays of string pointers cause lots of
1777 relocations when a shared library is dynamically loaded. A technique of
1778 using a single long string with a table of offsets can drastically reduce
1779 these. I have refactored PCRE in four places to do this. The result is
1780 dramatic:
1781
1782 Originally: 290
1783 After changing UCP table: 187
1784 After changing error message table: 43
1785 After changing table of "verbs" 36
1786 After changing table of Posix names 22
1787
1788 Thanks to the folks working on Gregex for glib for this insight.
1789
1790 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1791 unicode-properties was also set.
1792
1793 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1794
1795 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1796 checked only for CRLF.
1797
1798 14. Added casts to pcretest.c to avoid compiler warnings.
1799
1800 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1801
1802 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1803 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1804 entirely. This removes changes made in 7 above.
1805
1806 17. The CMake files have been updated, and there is now more information about
1807 building with CMake in the NON-UNIX-USE document.
1808
1809
1810 Version 7.3 28-Aug-07
1811 ---------------------
1812
1813 1. In the rejigging of the build system that eventually resulted in 7.1, the
1814 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1815 brackets there is not right, since it causes compilers to look for an
1816 installed pcre.h, not the version that is in the source that is being
1817 compiled (which of course may be different). I have changed it back to:
1818
1819 #include "pcre.h"
1820
1821 I have a vague recollection that the change was concerned with compiling in
1822 different directories, but in the new build system, that is taken care of
1823 by the VPATH setting the Makefile.
1824
1825 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1826 when the subject happened to end in the byte 0x85 (e.g. if the last
1827 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1828 characters but of course it shouldn't be taken as a newline when it is part
1829 of another character. The bug was that, for an unlimited repeat of . in
1830 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1831 characters when looking for a newline.
1832
1833 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1834
1835 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1836 in debug output.
1837
1838 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1839 long printing in the pcrecpp unittest when running under MinGW.
1840
1841 6. ESC_K was left out of the EBCDIC table.
1842
1843 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1844 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1845 limit also applies to "virtual nesting" when a pattern is recursive, and in
1846 this case 1000 isn't so big. I have been able to remove this limit at the
1847 expense of backing off one optimization in certain circumstances. Normally,
1848 when pcre_exec() would call its internal match() function recursively and
1849 immediately return the result unconditionally, it uses a "tail recursion"
1850 feature to save stack. However, when a subpattern that can match an empty
1851 string has an unlimited repetition quantifier, it no longer makes this
1852 optimization. That gives it a stack frame in which to save the data for
1853 checking that an empty string has been matched. Previously this was taken
1854 from the 1000-entry workspace that had been reserved. So now there is no
1855 explicit limit, but more stack is used.
1856
1857 8. Applied Daniel's patches to solve problems with the import/export magic
1858 syntax that is required for Windows, and which was going wrong for the
1859 pcreposix and pcrecpp parts of the library. These were overlooked when this
1860 problem was solved for the main library.
1861
1862 9. There were some crude static tests to avoid integer overflow when computing
1863 the size of patterns that contain repeated groups with explicit upper
1864 limits. As the maximum quantifier is 65535, the maximum group length was
1865 set at 30,000 so that the product of these two numbers did not overflow a
1866 32-bit integer. However, it turns out that people want to use groups that
1867 are longer than 30,000 bytes (though not repeat them that many times).
1868 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1869 made it possible to implement the integer overflow checks in a much more
1870 dynamic way, which I have now done. The artificial limitation on group
1871 length has been removed - we now have only the limit on the total length of
1872 the compiled pattern, which depends on the LINK_SIZE setting.
1873
1874 10. Fixed a bug in the documentation for get/copy named substring when
1875 duplicate names are permitted. If none of the named substrings are set, the
1876 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1877 empty string.
1878
1879 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1880 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1881 because the ] is interpreted as the first data character and the
1882 terminating ] is not found. PCRE has been made compatible with Perl in this
1883 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1884 cause memory overwriting.
1885
1886 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1887 string has been matched (to stop an infinite loop). It was not recognizing
1888 a conditional subpattern that could match an empty string if that
1889 subpattern was within another subpattern. For example, it looped when
1890 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1891 condition was not nested. This bug has been fixed.
1892
1893 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1894 past the start of the subject in the presence of bytes with the top bit
1895 set, for example "\x8aBCD".
1896
1897 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1898 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1899
1900 14. Optimized (?!) to (*FAIL).
1901
1902 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1903 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1904 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1905 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1906 does: it's just the validity check that is more restrictive.
1907
1908 16. Inserted checks for integer overflows during escape sequence (backslash)
1909 processing, and also fixed erroneous offset values for syntax errors during
1910 backslash processing.
1911
1912 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1913 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1914
1915 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1916 caused an overrun.
1917
1918 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1919 something other than just ASCII characters) inside a group that had an
1920 unlimited repeat caused a loop at compile time (while checking to see
1921 whether the group could match an empty string).
1922
1923 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1924 [\P{Any}] did so. (Error in the code for printing property names.)
1925
1926 21. An orphan \E inside a character class could cause a crash.
1927
1928 22. A repeated capturing bracket such as (A)? could cause a wild memory
1929 reference during compilation.
1930
1931 23. There are several functions in pcre_compile() that scan along a compiled
1932 expression for various reasons (e.g. to see if it's fixed length for look
1933 behind). There were bugs in these functions when a repeated \p or \P was
1934 present in the pattern. These operators have additional parameters compared
1935 with \d, etc, and these were not being taken into account when moving along
1936 the compiled data. Specifically:
1937
1938 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1939 length.
1940
1941 (b) An item such as \pL+ within a repeated group could cause crashes or
1942 loops.
1943
1944 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1945 "reference to non-existent subpattern" error.
1946
1947 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1948
1949 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1950 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1951
1952 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1953
1954 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1955 character were causing crashes (broken optimization).
1956
1957 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1958 \p or \P) caused a compile-time loop.
1959
1960 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1961 break. For example, the unstudied pattern [\r\n]A does not match the string
1962 "\r\nA" because change 7.0/46 below moves the current point on by two
1963 characters after failing to match at the start. However, the pattern \nA
1964 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1965 the same is true. There doesn't seem any very clean way out of this, but
1966 what I have chosen to do makes the common cases work: PCRE now takes note
1967 of whether there can be an explicit match for \r or \n anywhere in the
1968 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1969 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1970 pattern has explicit CR or LF references.
1971
1972 29. Added (*CR) etc for changing newline setting at start of pattern.
1973
1974
1975 Version 7.2 19-Jun-07
1976 ---------------------
1977
1978 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1979 which is apparently normally available under Windows.
1980
1981 2. Re-jig the pcregrep tests with different newline settings in an attempt
1982 to make them independent of the local environment's newline setting.
1983
1984 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1985
1986 4. Some of the "internals" tests were previously cut out when the link size
1987 was not 2, because the output contained actual offsets. The recent new
1988 "Z" feature of pcretest means that these can be cut out, making the tests
1989 usable with all link sizes.
1990
1991 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1992 stack recursion. This gives a massive performance boost under BSD, but just
1993 a small improvement under Linux. However, it saves one field in the frame
1994 in all cases.
1995
1996 6. Added more features from the forthcoming Perl 5.10:
1997
1998 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1999 recursion call. It refers to the nth most recently opened parentheses.
2000
2001 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2002 to be opened parentheses.
2003
2004 (c) Conditions that refer to capturing parentheses can be specified
2005 relatively, for example, (?(-2)... or (?(+3)...
2006
2007 (d) \K resets the start of the current match so that everything before
2008 is not part of it.
2009
2010 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2011
2012 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2013 reference syntax.
2014
2015 (g) (?| introduces a group in which the numbering of parentheses in each
2016 alternative starts with the same number.
2017
2018 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2019
2020 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2021 PCRE_INFO_JCHANGED.
2022
2023 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2024 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2025 for detecting groups that can match an empty string.
2026
2027 9. A pattern with a very large number of alternatives (more than several
2028 hundred) was running out of internal workspace during the pre-compile
2029 phase, where pcre_compile() figures out how much memory will be needed. A
2030 bit of new cunning has reduced the workspace needed for groups with
2031 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2032 workspace instead of running out of the 4096 that are available.
2033
2034 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2035
2036 11. Applied patch from Google to remove an optimization that didn't quite work.
2037 The report of the bug said:
2038
2039 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2040 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2041 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2042
2043 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2044 it matched the wrong number of bytes.
2045
2046
2047 Version 7.1 24-Apr-07
2048 ---------------------
2049
2050 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2051 that is more "standard", making use of automake and other Autotools. There
2052 is some re-arrangement of the files and adjustment of comments consequent
2053 on this.
2054
2055 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2056 for recursive directory scanning broke on some systems because the files
2057 are not scanned in any specific order and on different systems the order
2058 was different. A call to "sort" has been inserted into RunGrepTest for the
2059 approprate test as a short-term fix. In the longer term there may be an
2060 alternative.
2061
2062 3. I had an email from Eric Raymond about problems translating some of PCRE's
2063 man pages to HTML (despite the fact that I distribute HTML pages, some
2064 people do their own conversions for various reasons). The problems
2065 concerned the use of low-level troff macros .br and .in. I have therefore
2066 removed all such uses from the man pages (some were redundant, some could
2067 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2068 HTML has been updated to handle .nf/.fi and to complain if it encounters
2069 .br or .in.
2070
2071 4. Updated comments in configure.ac that get placed in config.h.in and also
2072 arranged for config.h to be included in the distribution, with the name
2073 config.h.generic, for the benefit of those who have to compile without
2074 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2075
2076 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2077 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2078 makevp.bat for the current PCRE, using the additional files
2079 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2080
2081 6. A Windows user reported a minor discrepancy with test 2, which turned out
2082 to be caused by a trailing space on an input line that had got lost in his
2083 copy. The trailing space was an accident, so I've just removed it.
2084
2085 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2086 that is needed.
2087
2088 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2089 as "const" (a) because they are and (b) because it helps the PHP
2090 maintainers who have recently made a script to detect big data structures
2091 in the php code that should be moved to the .rodata section. I remembered
2092 to update Builducptable as well, so it won't revert if ucptable.h is ever
2093 re-created.
2094
2095 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2096 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2097 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2098 support is not required. This saves 1.5-2K of code, which is important in
2099 some applications.
2100
2101 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2102 so as not to refer to the tables, even though these functions will never be
2103 called when UTF-8 support is disabled. Otherwise there are problems with a
2104 shared library.
2105
2106 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2107
2108 (a) It was defining its arguments as char * instead of void *.
2109
2110 (b) It was assuming that all moves were upwards in memory; this was true
2111 a long time ago when I wrote it, but is no longer the case.
2112
2113 The emulated memove() is provided for those environments that have neither
2114 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2115 is clearly not the case, as these two bugs were recently reported.
2116
2117 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2118 and Detrail to create the HTML documentation, the .txt form of the man
2119 pages, and it removes trailing spaces from listed files. It also creates
2120 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2121 case, it wraps all the #defines with #ifndefs. This script should be run
2122 before "make dist".
2123
2124 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2125 with Unicode property support.
2126
2127 (a) For a maximizing quantifier, if the two different cases of the
2128 character were of different lengths in their UTF-8 codings (there are
2129 some cases like this - I found 11), and the matching function had to
2130 back up over a mixture of the two cases, it incorrectly assumed they
2131 were both the same length.
2132
2133 (b) When PCRE was configured to use the heap rather than the stack for
2134 recursion during matching, it was not correctly preserving the data for
2135 the other case of a UTF-8 character when checking ahead for a match
2136 while processing a minimizing repeat. If the check also involved
2137 matching a wide character, but failed, corruption could cause an
2138 erroneous result when trying to check for a repeat of the original
2139 character.
2140
2141 13. Some tidying changes to the testing mechanism:
2142
2143 (a) The RunTest script now detects the internal link size and whether there
2144 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2145 values substituted by "configure". (The RunGrepTest script already did
2146 this for UTF-8.) The configure.ac script no longer substitutes the
2147 relevant variables.
2148
2149 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2150 with length and offset values. This means that the output is different
2151 for different internal link sizes. Test 2 is skipped for link sizes
2152 other than 2 because of this, bypassing the problem. Unfortunately,
2153 there was also a test in test 3 (the locale tests) that used /B and
2154 failed for link sizes other than 2. Rather than cut the whole test out,
2155 I have added a new /Z option to pcretest that replaces the length and
2156 offset values with spaces. This is now used to make test 3 independent
2157 of link size. (Test 2 will be tidied up later.)
2158
2159 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2160 segmentation fault instead of returning the appropriate error message.
2161
2162 15. In multiline mode when the newline sequence was set to "any", the pattern
2163 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2164 This doesn't seem right; it now treats the CRLF combination as the line
2165 ending, and so does not match in that case. It's only a pattern such as ^$
2166 that would hit this one: something like ^ABC$ would have failed after \r
2167 and then tried again after \r\n.
2168
2169 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2170 in an attempt to make files that differ only in their line terminators
2171 compare equal. This works on Linux.
2172
2173 17. Under certain error circumstances pcregrep might try to free random memory
2174 as it exited. This is now fixed, thanks to valgrind.
2175
2176 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2177 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2178 was because its rules for how to advance for /g after matching an empty
2179 string at the end of a line did not allow for this case. They now check for
2180 it specially.
2181
2182 20. pcretest is supposed to handle patterns and data of any length, by
2183 extending its buffers when necessary. It was getting this wrong when the
2184 buffer for a data line had to be extended.
2185
2186 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2187 CRLF as a newline sequence.
2188
2189 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2190 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2191 I have nevertheless tidied it up.
2192
2193 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2194
2195 24. Added a man page for pcre-config.
2196
2197
2198 Version 7.0 19-Dec-06
2199 ---------------------
2200
2201 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2202 moving to gcc 4.1.1.
2203
2204 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2205 sys/time.h, which is documented as needed for this function. It doesn't
2206 seem to matter on Linux, but it showed up on some releases of OS X.
2207
2208 3. It seems that there are systems where bytes whose values are greater than
2209 127 match isprint() in the "C" locale. The "C" locale should be the
2210 default when a C program starts up. In most systems, only ASCII printing
2211 characters match isprint(). This difference caused the output from pcretest
2212 to vary, making some of the tests fail. I have changed pcretest so that:
2213
2214 (a) When it is outputting text in the compiled version of a pattern, bytes
2215 other than 32-126 are always shown as hex escapes.
2216
2217 (b) When it is outputting text that is a matched part of a subject string,
2218 it does the same, unless a different locale has been set for the match
2219 (using the /L modifier). In this case, it uses isprint() to decide.
2220
2221 4. Fixed a major bug that caused incorrect computation of the amount of memory
2222 required for a compiled pattern when options that changed within the
2223 pattern affected the logic of the preliminary scan that determines the
2224 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2225 that the computed length was too small. The symptoms of this bug were
2226 either the PCRE error "internal error: code overflow" from pcre_compile(),
2227 or a glibc crash with a message such as "pcretest: free(): invalid next
2228 size (fast)". Examples of patterns that provoked this bug (shown in
2229 pcretest format) are:
2230
2231 /(?-x: )/x
2232 /(?x)(?-x: \s*#\s*)/
2233 /((?i)[\x{c0}])/8
2234 /(?i:[\x{c0}])/8
2235
2236 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2237 is now done differently.
2238
2239 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2240 wrapper classes; (b) implement a new function in the C++ scanner that is
2241 more efficient than the old way of doing things because it avoids levels of
2242 recursion in the regex matching; (c) add a paragraph to the documentation
2243 for the FullMatch() function.
2244
2245 6. The escape sequence \n was being treated as whatever was defined as
2246 "newline". Not only was this contrary to the documentation, which states
2247 that \n is character 10 (hex 0A), but it also went horribly wrong when
2248 "newline" was defined as CRLF. This has been fixed.
2249
2250 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2251 was being set to -1 for the "end of line" case (supposedly a value that no
2252 character can have). Though this value is never used (the check for end of
2253 line is "zero bytes in current character"), it caused compiler complaints.
2254 I've changed it to 0xffffffff.
2255
2256 8. In pcre_version.c, the version string was being built by a sequence of
2257 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2258 string (as it is for production releases) called a macro with an empty
2259 argument. The C standard says the result of this is undefined. The gcc
2260 compiler treats it as an empty string (which was what was wanted) but it is
2261 reported that Visual C gives an error. The source has been hacked around to
2262 avoid this problem.
2263
2264 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2265 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2266 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2267 of them did).
2268
2269 10. Originally, pcretest opened its input and output without "b"; then I was
2270 told that "b" was needed in some environments, so it was added for release
2271 5.0 to both the input and output. (It makes no difference on Unix-like
2272 systems.) Later I was told that it is wrong for the input on Windows. I've
2273 now abstracted the modes into two macros, to make it easier to fiddle with
2274 them, and removed "b" from the input mode under Windows.
2275
2276 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2277
2278 12. Added -help and --help to pcretest as an official way of being reminded
2279 of the options.
2280
2281 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2282 and pcrecpp.cc because they annoy compilers at high warning levels.
2283
2284 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2285
2286 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2287 scripts are not C programs :-) and which was not noticed because it works
2288 on Linux.
2289
2290 16. pcretest is supposed to handle any length of pattern and data line (as one
2291 line or as a continued sequence of lines) by extending its input buffer if
2292 necessary. This feature was broken for very long pattern lines, leading to
2293 a string of junk being passed to pcre_compile() if the pattern was longer
2294 than about 50K.
2295
2296 17. I have done a major re-factoring of the way pcre_compile() computes the
2297 amount of memory needed for a compiled pattern. Previously, there was code
2298 that made a preliminary scan of the pattern in order to do this. That was
2299 OK when PCRE was new, but as the facilities have expanded, it has become
2300 harder and harder to keep it in step with the real compile phase, and there
2301 have been a number of bugs (see for example, 4 above). I have now found a
2302 cunning way of running the real compile function in a "fake" mode that
2303 enables it to compute how much memory it would need, while actually only
2304 ever using a few hundred bytes of working memory and without too many
2305 tests of the mode. This should make future maintenance and development
2306 easier. A side effect of this work is that the limit of 200 on the nesting
2307 depth of parentheses has been removed (though this was never a serious
2308 limitation, I suspect). However, there is a downside: pcre_compile() now
2309 runs more slowly than before (30% or more, depending on the pattern). I
2310 hope this isn't a big issue. There is no effect on runtime performance.
2311
2312 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2313 newline (only possible for the last line of a file) and it was a
2314 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2315
2316 19. Added additional timing features to pcretest. (1) The -tm option now times
2317 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2318 separate command line item, by a number that specifies the number of
2319 repeats to use when timing. The default is 50000; this gives better
2320 precision, but takes uncomfortably long for very large patterns.
2321
2322 20. Extended pcre_study() to be more clever in cases where a branch of a
2323 subpattern has no definite first character. For example, (a*|b*)[cd] would
2324 previously give no result from pcre_study(). Now it recognizes that the
2325 first character must be a, b, c, or d.
2326
2327 21. There was an incorrect error "recursive call could loop indefinitely" if
2328 a subpattern (or the entire pattern) that was being tested for matching an
2329 empty string contained only one non-empty item after a nested subpattern.
2330 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2331 incorrectly, because the \d was being skipped in the check.
2332
2333 22. The pcretest program now has a new pattern option /B and a command line
2334 option -b, which is equivalent to adding /B to every pattern. This causes
2335 it to show the compiled bytecode, without the additional information that
2336 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2337 is the same as /B/I).
2338
2339 23. A new optimization is now able automatically to treat some sequences such
2340 as a*b as a*+b. More specifically, if something simple (such as a character
2341 or a simple class like \d) has an unlimited quantifier, and is followed by
2342 something that cannot possibly match the quantified thing, the quantifier
2343 is automatically "possessified".
2344
2345 24. A recursive reference to a subpattern whose number was greater than 39
2346 went wrong under certain circumstances in UTF-8 mode. This bug could also
2347 have affected the operation of pcre_study().
2348
2349 25. Realized that a little bit of performance could be had by replacing
2350 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2351
2352 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2353
2354 27. Possessive quantifiers such as a++ were previously implemented by turning
2355 them into atomic groups such as ($>a+). Now they have their own opcodes,
2356 which improves performance. This includes the automatically created ones
2357 from 23 above.
2358
2359 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2360 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2361 the first matched character to be a colon. This applied both to named and
2362 numbered groups.
2363
2364 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2365
2366 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2367 building PCRE on Windows easier, so I have included it in the distribution.
2368
2369 31. There is now a check in pcretest against a ridiculously large number being
2370 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2371 loop, the loop is abandoned.
2372
2373 32. Forward references to subpatterns in conditions such as (?(2)...) where
2374 subpattern 2 is defined later cause pcre_compile() to search forwards in
2375 the pattern for the relevant set of parentheses. This search went wrong
2376 when there were unescaped parentheses in a character class, parentheses
2377 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2378
2379 33. "Subroutine" calls and backreferences were previously restricted to
2380 referencing subpatterns earlier in the regex. This restriction has now
2381 been removed.
2382
2383 34. Added a number of extra features that are going to be in Perl 5.10. On the
2384 whole, these are just syntactic alternatives for features that PCRE had
2385 previously implemented using the Python syntax or my own invention. The
2386 other formats are all retained for compatibility.
2387
2388 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2389 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2390 also .NET compatible.
2391
2392 (b) A recursion or subroutine call to a named group can now be defined as
2393 (?&name) as well as (?P>name).
2394
2395 (c) A backreference to a named group can now be defined as \k<name> or
2396 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2397 5.10, are also .NET compatible.
2398
2399 (d) A conditional reference to a named group can now use the syntax
2400 (?(<name>) or (?('name') as well as (?(name).
2401
2402 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2403 groups (named and numbered) that are never evaluated inline, but can be
2404 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2405 is always false. There may be only one alternative in such a group.
2406
2407 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2408 as the simple (?(R). The condition is true only if the most recent
2409 recursion is that of the given number or name. It does not search out
2410 through the entire recursion stack.
2411
2412 (g) The escape \gN or \g{N} has been added, where N is a positive or
2413 negative number, specifying an absolute or relative reference.
2414
2415 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2416 some "unreachable code" warnings.
2417
2418 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2419 things, this adds five new scripts.
2420
2421 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2422 There were also incompatibilities regarding the handling of \Q..\E inside
2423 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2424 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2425
2426 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2427 matches an empty string, and forcibly breaks the loop. There were bugs in
2428 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2429 against aaaa the result was just "a" rather than "aaaa", for example. Two
2430 separate and independent bugs (that affected different cases) have been
2431 fixed.
2432
2433 39. Refactored the code to abolish the use of different opcodes for small
2434 capturing bracket numbers. This is a tidy that I avoided doing when I
2435 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2436 The new approach is not only tidier, it makes it possible to reduce the
2437 memory needed to fix the previous bug (38).
2438
2439 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2440 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2441 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2442 mode.
2443
2444 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2445 report.
2446
2447 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2448 copy construction and assignment in the C++ wrapper.
2449
2450 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2451 couple of bugs that could have given wrong results in the "--newline=crlf"
2452 case.
2453
2454 44. Added a number of casts and did some reorganization of signed/unsigned int
2455 variables following suggestions from Dair Grant. Also renamed the variable
2456 "this" as "item" because it is a C++ keyword.
2457
2458 45. Arranged for dftables to add
2459
2460 #include "pcre_internal.h"
2461
2462 to pcre_chartables.c because without it, gcc 4.x may remove the array
2463 definition from the final binary if PCRE is built into a static library and
2464 dead code stripping is activated.
2465
2466 46. For an unanchored pattern, if a match attempt fails at the start of a
2467 newline sequence, and the newline setting is CRLF or ANY, and the next two
2468 characters are CRLF, advance by two characters instead of one.
2469
2470
2471 Version 6.7 04-Jul-06
2472 ---------------------
2473
2474 1. In order to handle tests when input lines are enormously long, pcretest has
2475 been re-factored so that it automatically extends its buffers when
2476 necessary. The code is crude, but this _is_ just a test program. The
2477 default size has been increased from 32K to 50K.
2478
2479 2. The code in pcre_study() was using the value of the re argument before
2480 testing it for NULL. (Of course, in any sensible call of the function, it
2481 won't be NULL.)
2482
2483 3. The memmove() emulation function in pcre_internal.h, which is used on
2484 systems that lack both memmove() and bcopy() - that is, hardly ever -
2485 was missing a "static" storage class specifier.
2486
2487 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2488 containing an extended class (one that cannot be represented by a bitmap
2489 because it contains high-valued characters or Unicode property items, e.g.
2490 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2491 pattern, but PCRE should not loop if you do not (it no longer does).
2492 [Detail: two cases were found: (a) a repeated subpattern containing an
2493 extended class; (b) a recursive reference to a subpattern that followed a
2494 previous extended class. It wasn't skipping over the extended class
2495 correctly when UTF-8 mode was not set.]
2496
2497 5. A negated single-character class was not being recognized as fixed-length
2498 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2499 compile error "lookbehind assertion is not fixed length".
2500
2501 6. The RunPerlTest auxiliary script was showing an unexpected difference
2502 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2503 write a Perl script that can interpret lines of an input file either as
2504 byte characters or as UTF-8, which is what "perltest" was being required to
2505 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2506 can't do is switch easily at run time between having the "use utf8;" pragma
2507 or not. In the end, I fudged it by using the RunPerlTest script to insert
2508 "use utf8;" explicitly for the UTF-8 tests.
2509
2510 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2511 the end of the subject string, contrary to the documentation and to what
2512 Perl does. This was true of both matching functions. Now it matches only at
2513 the start of the subject and immediately after *internal* newlines.
2514
2515 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2516 a pointer to an int instead of a pointer to an unsigned long int. This
2517 caused problems on 64-bit systems.
2518
2519 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2520 instance of the 'standard' template library not being so standard".
2521
2522 10. There was no check on the number of named subpatterns nor the maximum
2523 length of a subpattern name. The product of these values is used to compute
2524 the size of the memory block for a compiled pattern. By supplying a very
2525 long subpattern name and a large number of named subpatterns, the size
2526 computation could be caused to overflow. This is now prevented by limiting
2527 the length of names to 32 characters, and the number of named subpatterns
2528 to 10,000.
2529
2530 11. Subpatterns that are repeated with specific counts have to be replicated in
2531 the compiled pattern. The size of memory for this was computed from the
2532 length of the subpattern and the repeat count. The latter is limited to
2533 65535, but there was no limit on the former, meaning that integer overflow
2534 could in principle occur. The compiled length of a repeated subpattern is
2535 now limited to 30,000 bytes in order to prevent this.
2536
2537 12. Added the optional facility to have named substrings with the same name.
2538
2539 13. Added the ability to use a named substring as a condition, using the
2540 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2541 are numbers (not recommended). Forward references are permitted.
2542
2543 14. Added forward references in named backreferences (if you see what I mean).
2544
2545 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2546 pattern could run off the end of the subject. For example, the pattern
2547 "(?s)(.{1,5})"8 did this with the subject "ab".
2548
2549 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2550 PCRE_CASELESS was set when matching characters that were quantified with ?
2551 or *.
2552
2553 17. A character class other than a single negated character that had a minimum
2554 but no maximum quantifier - for example [ab]{6,} - was not handled
2555 correctly by pce_dfa_exec(). It would match only one character.
2556
2557 18. A valid (though odd) pattern that looked like a POSIX character
2558 class but used an invalid character after [ (for example [[,abc,]]) caused
2559 pcre_compile() to give the error "Failed: internal error: code overflow" or
2560 in some cases to crash with a glibc free() error. This could even happen if
2561 the pattern terminated after [[ but there just happened to be a sequence of
2562 letters, a binary zero, and a closing ] in the memory that followed.
2563
2564 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2565 over the years. Originally (before any Unicode support), just the bottom 8
2566 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2567 output from "man perlunicode" includes this:
2568
2569 The regular expression compiler produces polymorphic opcodes. That
2570 is, the pattern adapts to the data and automatically switches to
2571 the Unicode character scheme when presented with Unicode data--or
2572 instead uses a traditional byte scheme when presented with byte
2573 data.
2574
2575 Sadly, a wide octal escape does not cause a switch, and in a string with
2576 no other multibyte characters, these octal escapes are treated as before.
2577 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2578 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2579 Unicode string.
2580
2581 I have not perpetrated such confusion in PCRE. Up till now, it took just
2582 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2583 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2584 translate to the appropriate multibyte character.
2585
2586 29. Applied some refactoring to reduce the number of warnings from Microsoft
2587 and Borland compilers. This has included removing the fudge introduced
2588 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2589 a warning about an unused variable.
2590
2591 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2592 characters since release 4.0, because Perl (from release 5.004) does not.
2593 [Or at least, is documented not to: some releases seem to be in conflict
2594 with the documentation.] However, when a pattern was studied with
2595 pcre_study() and all its branches started with \s, PCRE still included VT
2596 as a possible starting character. Of course, this did no harm; it just
2597 caused an unnecessary match attempt.
2598
2599 22. Removed a now-redundant internal flag bit that recorded the fact that case
2600 dependency changed within the pattern. This was once needed for "required
2601 byte" processing, but is no longer used. This recovers a now-scarce options
2602 bit. Also moved the least significant internal flag bit to the most-
2603 significant bit of the word, which was not previously used (hangover from
2604 the days when it was an int rather than a uint) to free up another bit for
2605 the future.
2606
2607 23. Added support for CRLF line endings as well as CR and LF. As well as the
2608 default being selectable at build time, it can now be changed at runtime
2609 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2610 specify that it is scanning data with non-default line endings.
2611
2612 24. Changed the definition of CXXLINK to make it agree with the definition of
2613 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2614
2615 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2616 recursions. This makes a big different to stack usage for some patterns.
2617
2618 26. If a subpattern containing a named recursion or subroutine reference such
2619 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2620 the space required for the compiled pattern went wrong and gave too small a
2621 value. Depending on the environment, this could lead to "Failed: internal
2622 error: code overflow at offset 49" or "glibc detected double free or
2623 corruption" errors.
2624
2625 27. Applied patches from Google (a) to support the new newline modes and (b) to
2626 advance over multibyte UTF-8 characters in GlobalReplace.
2627
2628 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2629 difference for some implementation of PCRE in some Windows version.
2630
2631 29. Added some extra testing facilities to pcretest:
2632
2633 \q<number> in a data line sets the "match limit" value
2634 \Q<number> in a data line sets the "match recursion limt" value
2635 -S <number> sets the stack size, where <number> is in megabytes
2636
2637 The -S option isn't available for Windows.
2638
2639
2640 Version 6.6 06-Feb-06
2641 ---------------------
2642
2643 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2644 in pcreposix.h. I have copied the definition from pcre.h.
2645
2646 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2647 because pcre.h is no longer a built file.
2648
2649 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2650 not normally included in the compiled code.
2651
2652
2653 Version 6.5 01-Feb-06
2654 ---------------------
2655
2656 1. When using the partial match feature with pcre_dfa_exec(), it was not
2657 anchoring the second and subsequent partial matches at the new starting
2658 point. This could lead to incorrect results. For example, with the pattern
2659 /1234/, partially matching against "123" and then "a4" gave a match.
2660
2661 2. Changes to pcregrep:
2662
2663 (a) All non-match returns from pcre_exec() were being treated as failures
2664 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2665 error message is output. Some extra information is given for the
2666 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2667 probably the only errors that are likely to be caused by users (by
2668 specifying a regex that has nested indefinite repeats, for instance).
2669 If there are more than 20 of these errors, pcregrep is abandoned.
2670
2671 (b) A binary zero was treated as data while matching, but terminated the
2672 output line if it was written out. This has been fixed: binary zeroes
2673 are now no different to any other data bytes.
2674
2675 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2676 used to set a locale for matching. The --locale=xxxx long option has
2677 been added (no short equivalent) to specify a locale explicitly on the
2678 pcregrep command, overriding the environment variables.
2679
2680 (d) When -B was used with -n, some line numbers in the output were one less
2681 than they should have been.
2682
2683 (e) Added the -o (--only-matching) option.
2684
2685 (f) If -A or -C was used with -c (count only), some lines of context were
2686 accidentally printed for the final match.
2687
2688 (g) Added the -H (--with-filename) option.
2689
2690 (h) The combination of options -rh failed to suppress file names for files
2691 that were found from directory arguments.
2692
2693 (i) Added the -D (--devices) and -d (--directories) options.
2694
2695 (j) Added the -F (--fixed-strings) option.
2696
2697 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2698
2699 (l) Added the --colo(u)r option.
2700
2701 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2702 is not present by default.
2703
2704 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2705 items such as (?R) or (?1), when the recursion could match a number of
2706 alternatives. If it matched one of the alternatives, but subsequently,
2707 outside the recursion, there was a failure, the code tried to back up into
2708 the recursion. However, because of the way PCRE is implemented, this is not
2709 possible, and the result was an incorrect result from the match.
2710
2711 In order to prevent this happening, the specification of recursion has
2712 been changed so that all such subpatterns are automatically treated as
2713 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2714
2715 4. I had overlooked the fact that, in some locales, there are characters for
2716 which isalpha() is true but neither isupper() nor islower() are true. In
2717 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2718 and ordfeminine) are like this. This affected the treatment of \w and \W
2719 when they appeared in character classes, but not when they appeared outside
2720 a character class. The bit map for "word" characters is now created
2721 separately from the results of isalnum() instead of just taking it from the
2722 upper, lower, and digit maps. (Plus the underscore character, of course.)
2723
2724 5. The above bug also affected the handling of POSIX character classes such as
2725 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2726 permanent tables. Instead, the bit maps for such a class were previously
2727 created as the appropriate unions of the upper, lower, and digit bitmaps.
2728 Now they are created by subtraction from the [[:word:]] class, which has
2729 its own bitmap.
2730
2731 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2732 It is created by subtracting the vertical space characters (\x09, \x0a,
2733 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2734 subtraction was done in the overall bitmap for a character class, meaning
2735 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2736 be recognized. This bug has been fixed.
2737
2738 7. Patches from the folks at Google:
2739
2740 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2741 real life, but is still worth protecting against".
2742
2743 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2744 regular expressions".
2745
2746 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2747 have it.
2748
2749 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2750 "configure" and the latter not, in order to fix a problem somebody had
2751 with compiling the Arg class on HP-UX.
2752
2753 (e) Improve the error-handling of the C++ wrapper a little bit.
2754
2755 (f) New tests for checking recursion limiting.
2756
2757 8. The pcre_memmove() function, which is used only if the environment does not
2758 have a standard memmove() function (and is therefore rarely compiled),
2759 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2760 returning a result (though PCRE never actually uses the result).
2761
2762 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2763 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2764 returned instead of calling malloc() with an overflowing number that would
2765 most likely cause subsequent chaos.
2766
2767 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2768
2769 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2770 with this option is matched, the nmatch and pmatch options of regexec() are
2771 ignored.
2772
2773 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2774 provided in case anyone wants to the the POSIX interface with UTF-8
2775 strings.
2776
2777 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2778 C++ linking (needed for some HP-UX environments).
2779
2780 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2781 (unused parameter) and in the pcre_printint() function (omitted "default"
2782 switch label when the default is to do nothing).
2783
2784 15. Added some code to make it possible, when PCRE is compiled as a C++
2785 library, to replace subject pointers for pcre_exec() with a smart pointer
2786 class, thus making it possible to process discontinuous strings.
2787
2788 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2789 much the same function. They were added by different people who were trying
2790 to make PCRE easy to compile on non-Unix systems. It has been suggested
2791 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2792 for compiling on Windows systems. I have therefore replaced it with
2793 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2794 defaults to "extern" for C or "extern C" for C++, which works fine on
2795 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2796 SCOPE with something explicit in config.h. In addition:
2797
2798 (a) pcreposix.h still had just "extern" instead of either of these macros;
2799 I have replaced it with PCRE_DATA_SCOPE.
2800
2801 (b) Functions such as _pcre_xclass(), which are internal to the library,
2802 but external in the C sense, all had PCRE_EXPORT in their definitions.
2803 This is apparently wrong for the Windows case, so I have removed it.
2804 (It makes no difference on Unix-like systems.)
2805
2806 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2807 of recursive calls to match(). This is different to MATCH_LIMIT because
2808 that limits the total number of calls to match(), not all of which increase
2809 the depth of recursion. Limiting the recursion depth limits the amount of
2810 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2811 when PCRE is compiled, and changed at run time. A patch from Google adds
2812 this functionality to the C++ interface.
2813
2814 18. Changes to the handling of Unicode character properties:
2815
2816 (a) Updated the table to Unicode 4.1.0.
2817
2818 (b) Recognize characters that are not in the table as "Cn" (undefined).
2819
2820 (c) I revised the way the table is implemented to a much improved format
2821 which includes recognition of ranges. It now supports the ranges that
2822 are defined in UnicodeData.txt, and it also amalgamates other
2823 characters into ranges. This has reduced the number of entries in the
2824 table from around 16,000 to around 3,000, thus reducing its size
2825 considerably. I realized I did not need to use a tree structure after
2826 all - a binary chop search is just as efficient. Having reduced the
2827 number of entries, I extended their size from 6 bytes to 8 bytes to
2828 allow for more data.
2829
2830 (d) Added support for Unicode script names via properties such as \p{Han}.
2831
2832 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2833 matching that character.
2834
2835 20. When matching a repeated Unicode property with a minimum greater than zero,
2836 (for example \pL{2,}), PCRE could look past the end of the subject if it
2837 reached it while seeking the minimum number of characters. This could
2838 happen only if some of the characters were more than one byte long, because
2839 there is a check for at least the minimum number of bytes.
2840
2841 21. Refactored the implementation of \p and \P so as to be more general, to
2842 allow for more different types of property in future. This has changed the
2843 compiled form incompatibly. Anybody with saved compiled patterns that use
2844 \p or \P will have to recompile them.
2845
2846 22. Added "Any" and "L&" to the supported property types.
2847
2848 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2849 but give a compile time error if the value is greater than 0xff.
2850
2851 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2852 accidentally not being installed or uninstalled.
2853
2854 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2855 made were to insert the current release number. This seemed silly, because
2856 it made things harder for people building PCRE on systems that don't run
2857 "configure". I have turned pcre.h into a distributed file, no longer built
2858 by "configure", with the version identification directly included. There is
2859 no longer a pcre.h.in file.
2860
2861 However, this change necessitated a change to the pcre-config script as
2862 well. It is built from pcre-config.in, and one of the substitutions was the
2863 release number. I have updated configure.ac so that ./configure now finds
2864 the release number by grepping pcre.h.
2865
2866 26. Added the ability to run the tests under valgrind.
2867
2868
2869 Version 6.4 05-Sep-05
2870 ---------------------
2871
2872 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2873 "--" to be printed when multiple files were scanned, even when none of the
2874 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2875 consider it to be a bug, and have restored the previous behaviour.
2876
2877 2. A couple of code tidies to get rid of compiler warnings.
2878
2879 3. The pcretest program used to cheat by referring to symbols in the library
2880 whose names begin with _pcre_. These are internal symbols that are not
2881 really supposed to be visible externally, and in some environments it is
2882 possible to suppress them. The cheating is now confined to including
2883 certain files from the library's source, which is a bit cleaner.
2884
2885 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2886 file's purpose clearer.
2887
2888 5. Reorganized pcre_ucp_findchar().
2889
2890
2891 Version 6.3 15-Aug-05
2892 ---------------------
2893
2894 1. The file libpcre.pc.in did not have general read permission in the tarball.
2895
2896 2. There were some problems when building without C++ support:
2897
2898 (a) If C++ support was not built, "make install" and "make test" still
2899 tried to test it.
2900
2901 (b) There were problems when the value of CXX was explicitly set. Some
2902 changes have been made to try to fix these, and ...
2903
2904 (c) --disable-cpp can now be used to explicitly disable C++ support.
2905
2906 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2907 backslash in a target when C++ was disabled. This confuses some
2908 versions of "make", apparently. Using an intermediate variable solves
2909 this. (Same for CPP_LOBJ.)
2910
2911 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2912 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2913 necessary on certain architectures.
2914
2915 4. Added a setting of -export-symbols-regex to the link command to remove
2916 those symbols that are exported in the C sense, but actually are local
2917 within the library, and not documented. Their names all begin with
2918 "_pcre_". This is not a perfect job, because (a) we have to except some
2919 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2920 available (and never for static libraries). I have made a note to try to
2921 find a way round (a) in the future.
2922
2923
2924 Version 6.2 01-Aug-05
2925 ---------------------
2926
2927 1. There was no test for integer overflow of quantifier values. A construction
2928 such as {1111111111111111} would give undefined results. What is worse, if
2929 a minimum quantifier for a parenthesized subpattern overflowed and became
2930 negative, the calculation of the memory size went wrong. This could have
2931 led to memory overwriting.
2932
2933 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2934
2935 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2936 operating environments where this matters.
2937
2938 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2939 PCRE options from within the C++ wrapper.
2940
2941 5. Named capturing subpatterns were not being correctly counted when a pattern
2942 was compiled. This caused two problems: (a) If there were more than 100
2943 such subpatterns, the calculation of the memory needed for the whole
2944 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2945 back references of the form \12, where the number was greater than 9, were
2946 not recognized as back references, even though there were sufficient
2947 previous subpatterns.
2948
2949 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2950 versions of gcc, e.g. 2.95.4.
2951
2952
2953 Version 6.1 21-Jun-05
2954 ---------------------
2955
2956 1. There was one reference to the variable "posix" in pcretest.c that was not
2957 surrounded by "#if !defined NOPOSIX".
2958
2959 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2960 the cross-check on the old pcre_info() function, for the benefit of the
2961 cut-down version of PCRE that is currently imported into Exim.
2962
2963 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2964 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2965 patterns that start (?i) but I don't think that matters. The use of (?i) is
2966 just an example; this all applies to the other options as well.
2967
2968 4. Since libtool seems to echo the compile commands it is issuing, the output
2969 from "make" can be reduced a bit by putting "@" in front of each libtool
2970 compile command.
2971
2972 5. Patch from the folks at Google for configure.in to be a bit more thorough
2973 in checking for a suitable C++ installation before trying to compile the
2974 C++ stuff. This should fix a reported problem when a compiler was present,
2975 but no suitable headers.
2976
2977 6. The man pages all had just "PCRE" as their title. I have changed them to
2978 be the relevant file name. I have also arranged that these names are
2979 retained in the file doc/pcre.txt, which is a concatenation in text format
2980 of all the man pages except the little individual ones for each function.
2981
2982 7. The NON-UNIX-USE file had not been updated for the different set of source
2983 files that come with release 6. I also added a few comments about the C++
2984 wrapper.
2985
2986
2987 Version 6.0 07-Jun-05
2988 ---------------------
2989
2990 1. Some minor internal re-organization to help with my DFA experiments.
2991
2992 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2993 didn't matter for the library itself when fully configured, but did matter
2994 when compiling without UCP support, or within Exim, where the ucp files are
2995 not imported.
2996
2997 3. Refactoring of the library code to split up the various functions into
2998 different source modules. The addition of the new DFA matching code (see
2999 below) to a single monolithic source would have made it really too
3000 unwieldy, quite apart from causing all the code to be include in a
3001 statically linked application, when only some functions are used. This is
3002 relevant even without the DFA addition now that patterns can be compiled in
3003 one application and matched in another.
3004
3005 The downside of splitting up is that there have to be some external
3006 functions and data tables that are used internally in different modules of
3007 the library but which are not part of the API. These have all had their
3008 names changed to start with "_pcre_" so that they are unlikely to clash
3009 with other external names.
3010
3011 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3012 a different (DFA) algorithm. Although it is slower than the original
3013 function, it does have some advantages for certain types of matching
3014 problem.
3015
3016 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3017 including restarting after a partial match.
3018
3019 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3020 defined when compiling for Windows was sent to me. I have put it into the
3021 code, though I have no means of testing or verifying it.
3022
3023 7. Added the pcre_refcount() auxiliary function.
3024
3025 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3026 match before or at the first newline in the subject string. In pcretest,
3027 the /f option on a pattern can be used to set this.
3028
3029 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3030 would behave wrongly. This has been present in PCRE since release 4.0.
3031
3032 10. A number of changes to the pcregrep command:
3033
3034 (a) Refactored how -x works; insert ^(...)$ instead of setting
3035 PCRE_ANCHORED and checking the length, in preparation for adding
3036 something similar for -w.
3037
3038 (b) Added the -w (match as a word) option.
3039
3040 (c) Refactored the way lines are read and buffered so as to have more
3041 than one at a time available.
3042
3043 (d) Implemented a pcregrep test script.
3044
3045 (e) Added the -M (multiline match) option. This allows patterns to match
3046 over several lines of the subject. The buffering ensures that at least
3047 8K, or the rest of the document (whichever is the shorter) is available
3048 for matching (and similarly the previous 8K for lookbehind assertions).
3049
3050 (f) Changed the --help output so that it now says
3051
3052 -w, --word-regex(p)
3053
3054 instead of two lines, one with "regex" and the other with "regexp"
3055 because that confused at least one person since the short forms are the
3056 same. (This required a bit of code, as the output is generated
3057 automatically from a table. It wasn't just a text change.)
3058
3059 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3060 option but starts with a hyphen. Could be a pattern or a path name
3061 starting with a hyphen, for instance.
3062
3063 (h) "-" can be given as a file name to represent stdin.
3064
3065 (i) When file names are being printed, "(standard input)" is used for
3066 the standard input, for compatibility with GNU grep. Previously
3067 "<stdin>" was used.
3068
3069 (j) The option --label=xxx can be used to supply a name to be used for
3070 stdin when file names are being printed. There is no short form.
3071
3072 (k) Re-factored the options decoding logic because we are going to add
3073 two more options that take data. Such options can now be given in four
3074 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3075
3076 (l) Added the -A, -B, and -C options for requesting that lines of context
3077 around matches be printed.
3078
3079 (m) Added the -L option to print the names of files that do not contain
3080 any matching lines, that is, the complement of -l.
3081
3082 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3083 continue to scan other files.
3084
3085 (o) The -s option was incorrectly implemented. For compatibility with other
3086 greps, it now suppresses the error message for a non-existent or non-
3087 accessible file (but not the return code). There is a new option called
3088 -q that suppresses the output of matching lines, which was what -s was
3089 previously doing.
3090
3091 (p) Added --include and --exclude options to specify files for inclusion
3092 and exclusion when recursing.
3093
3094 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3095 Hopefully, it now does.
3096
3097 12. Missing cast in pcre_study().
3098
3099 13. Added an "uninstall" target to the makefile.
3100
3101 14. Replaced "extern" in the function prototypes in Makefile.in with
3102 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3103 world, but is set differently for Windows.
3104
3105 15. Added a second compiling function called pcre_compile2(). The only
3106 difference is that it has an extra argument, which is a pointer to an
3107 integer error code. When there is a compile-time failure, this is set
3108 non-zero, in addition to the error test pointer being set to point to an
3109 error message. The new argument may be NULL if no error number is required
3110 (but then you may as well call pcre_compile(), which is now just a
3111 wrapper). This facility is provided because some applications need a
3112 numeric error indication, but it has also enabled me to tidy up the way
3113 compile-time errors are handled in the POSIX wrapper.
3114
3115 16. Added VPATH=.libs to the makefile; this should help when building with one
3116 prefix path and installing with another. (Or so I'm told by someone who
3117 knows more about this stuff than I do.)
3118
3119 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3120 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3121 match everything, including newlines. This is not POSIX-compatible, but
3122 somebody wanted the feature. From pcretest it can be activated by using
3123 both the P and the s flags.
3124
3125 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3126
3127 19. libpcre.pc was being incorrectly installed as executable.
3128
3129 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3130 it now also looks for '\r' so that it will work unmodified on Windows.
3131
3132 21. Added Google's contributed C++ wrapper to the distribution.
3133
3134 22. Added some untidy missing memory free() calls in pcretest, to keep
3135 Electric Fence happy when testing.
3136
3137
3138
3139 Version 5.0 13-Sep-04
3140 ---------------------
3141
3142 1. Internal change: literal characters are no longer packed up into items
3143 containing multiple characters in a single byte-string. Each character
3144 is now matched using a separate opcode. However, there may be more than one
3145 byte in the character in UTF-8 mode.
3146
3147 2. The pcre_callout_block structure has two new fields: pattern_position and
3148 next_item_length. These contain the offset in the pattern to the next match
3149 item, and its length, respectively.
3150
3151 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3152 insertion of callouts before each pattern item. Added the /C option to
3153 pcretest to make use of this.
3154
3155 4. On the advice of a Windows user, the lines
3156
3157 #if defined(_WIN32) || defined(WIN32)
3158 _setmode( _fileno( stdout ), 0x8000 );
3159 #endif /* defined(_WIN32) || defined(WIN32) */
3160
3161 have been added to the source of pcretest. This apparently does useful
3162 magic in relation to line terminators.
3163
3164 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3165 for the benefit of those environments where the "b" makes a difference.
3166
3167 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3168 to know about it. I have put a hack into configure.in that adds in code
3169 to set GCC=yes if CC=icc. This seems to end up at a point in the
3170 generated configure script that is early enough to affect the setting of
3171 compiler options, which is what is needed, but I have no means of testing
3172 whether it really works. (The user who reported this had patched the
3173 generated configure script, which of course I cannot do.)
3174
3175 LATER: After change 22 below (new libtool files), the configure script
3176 seems to know about icc (and also ecc). Therefore, I have commented out
3177 this hack in configure.in.
3178
3179 7. Added support for pkg-config (2 patches were sent in).
3180
3181 8. Negated POSIX character classes that used a combination of internal tables
3182 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3183 [[:^ascii]]. Typically, they would match almost any characters. The other
3184 POSIX classes were not broken in this way.
3185
3186 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3187 to find the match, as PCRE was deluded into thinking that the match had to
3188 start at the start point or following a newline. The same bug applied to
3189 patterns with negative forward assertions or any backward assertions
3190 preceding ".*" at the start, unless the pattern required a fixed first
3191 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3192
3193 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3194 starting at the last subject character, bytes beyond the end of the subject
3195 string were read.
3196
3197 11. Renamed the variable "class" as "classbits" to make life easier for C++
3198 users. (Previously there was a macro definition, but it apparently wasn't
3199 enough.)
3200
3201 12. Added the new field "tables" to the extra data so that tables can be passed
3202 in at exec time, or the internal tables can be re-selected. This allows
3203 a compiled regex to be saved and re-used at a later time by a different
3204 program that might have everything at different addresses.
3205
3206 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3207 -R library as well as a -L library.
3208
3209 14. The debugging options of pcretest (-d on the command line or D on a
3210 pattern) showed incorrect output for anything following an extended class
3211 that contained multibyte characters and which was followed by a quantifier.
3212
3213 15. Added optional support for general category Unicode character properties
3214 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3215 support. It adds about 90K to the size of the library. The meanings of the
3216 inbuilt class escapes such as \d and \s have NOT been changed.
3217
3218 16. Updated pcredemo.c to include calls to free() to release the memory for the
3219 compiled pattern.
3220
3221 17. The generated file chartables.c was being created in the source directory
3222 instead of in the building directory. This caused the build to fail if the
3223 source directory was different from the building directory, and was
3224 read-only.
3225
3226 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3227 file. No doubt somebody will tell me if they don't make sense... Also added
3228 Dan Mooney's comments about building on OpenVMS.
3229
3230 19. Added support for partial matching via the PCRE_PARTIAL option for
3231 pcre_exec() and the \P data escape in pcretest.
3232
3233 20. Extended pcretest with 3 new pattern features:
3234
3235 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3236 write the compiled pattern to the file whose name is "rest-of-line".
3237 This is a straight binary dump of the data, with the saved pointer to
3238 the character tables forced to be NULL. The study data, if any, is
3239 written too. After writing, pcretest reads a new pattern.
3240
3241 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3242 compiled pattern from the given file. There must not be any
3243 occurrences of "<" in the file name (pretty unlikely); if there are,
3244 pcretest will instead treat the initial "<" as a pattern delimiter.
3245 After reading in the pattern, pcretest goes on to read data lines as
3246 usual.
3247
3248 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3249 and 16-bit fields in a compiled pattern, to simulate a pattern that
3250 was compiled on a host of opposite endianness.
3251
3252 21. The pcre-exec() function can now cope with patterns that were compiled on
3253 hosts of opposite endianness, with this restriction:
3254
3255 As for any compiled expression that is saved and used later, the tables
3256 pointer field cannot be preserved; the extra_data field in the arguments
3257 to pcre_exec() should be used to pass in a tables address if a value
3258 other than the default internal tables were used at compile time.
3259
3260 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3261 now diagnosed as an error. Previously, most of the time, a negative number
3262 would have been treated as zero, but if in addition "ovector" was passed as
3263 NULL, a crash could occur.
3264
3265 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3266 new versions from the libtool 1.5 distribution (the last one is a copy of
3267 a file called libtool.m4). This seems to have fixed the need to patch
3268 "configure" to support Darwin 1.3 (which I used to do). However, I still
3269 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3270 workstation).
3271
3272 24. Changed the PCRE licence to be the more standard "BSD" licence.
3273
3274
3275 Version 4.5 01-Dec-03
3276 ---------------------
3277
3278 1. There has been some re-arrangement of the code for the match() function so
3279 that it can be compiled in a version that does not call itself recursively.
3280 Instead, it keeps those local variables that need separate instances for
3281 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3282 needs to "recurse". Keeping track of where control must go is done by means
3283 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3284 hide most of the details from the main code, and operates only if
3285 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3286 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3287 operating.
3288
3289 To make it easier for callers to provide specially tailored get/free
3290 functions for this usage, two new functions, pcre_stack_malloc, and
3291 pcre_stack_free, are used. They are always called in strict stacking order,
3292 and the size of block requested is always the same.
3293
3294 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3295 PCRE has been compiled to use the stack or the heap for recursion. The
3296 -C option of pcretest uses this to show which version is compiled.
3297
3298 A new data escape \S, is added to pcretest; it causes the amounts of store
3299 obtained and freed by both kinds of malloc/free at match time to be added
3300 to the output.
3301
3302 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3303 what's available on my current Linux desktop machine.
3304
3305 3. When matching a UTF-8 string, the test for a valid string at the start has
3306 been extended. If start_offset is not zero, PCRE now checks that it points
3307 to a byte that is the start of a UTF-8 character. If not, it returns
3308 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3309 this is necessary because there may be backward assertions in the pattern.
3310 When matching the same subject several times, it may save resources to use
3311 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3312
3313 4. The code for checking the validity of UTF-8 strings has been tightened so
3314 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3315 containing "overlong sequences".
3316
3317 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3318 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3319 should have been "&", but it just so happened that all the cases this let
3320 through by mistake were picked up later in the function.
3321
3322 6. I had used a variable called "isblank" - this is a C99 function, causing
3323 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3324
3325 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3326 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3327 so the version is shown; (c) stop "make test" echoing "./RunTest".
3328
3329 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3330
3331 9. The prototype for memmove() for systems that don't have it was using
3332 size_t, but the inclusion of the header that defines size_t was later. I've
3333 moved the #includes for the C headers earlier to avoid this.
3334
3335 10. Added some adjustments to the code to make it easier to compiler on certain
3336 special systems:
3337
3338 (a) Some "const" qualifiers were missing.
3339 (b) Added the macro EXPORT before all exported functions; by default this
3340 is defined to be empty.
3341 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3342 that it reads its output file name as an argument instead of writing
3343 to the standard output and assuming this can be redirected.
3344
3345 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3346 class containing characters with values greater than 255, PCRE compilation
3347 went into a loop.
3348
3349 12. A recursive reference to a subpattern that was within another subpattern
3350 that had a minimum quantifier of zero caused PCRE to crash. For example,
3351 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3352 recursion. If the recursively-called subpattern itself had a zero repeat,
3353 that was OK.
3354
3355 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3356 buffer into which it was copied (for escape processing) was still set at
3357 1024, so long lines caused crashes.
3358
3359 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3360 "internal error: code overflow...". This applied to any character class
3361 that was followed by a possessive quantifier.
3362
3363 15. Modified the Makefile to add libpcre.la as a prerequisite for
3364 libpcreposix.la because I was told this is needed for a parallel build to
3365 work.
3366
3367 16. If a pattern that contained .* following optional items at the start was
3368 studied, the wrong optimizing data was generated, leading to matching
3369 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3370 matching string must start with a or b or c. The correct conclusion for
3371 this pattern is that a match can start with any character.
3372
3373
3374 Version 4.4 13-Aug-03
3375 ---------------------
3376
3377 1. In UTF-8 mode, a character class containing characters with values between
3378 127 and 255 was not handled correctly if the compiled pattern was studied.
3379 In fixing this, I have also improved the studying algorithm for such
3380 classes (slightly).
3381
3382 2. Three internal functions had redundant arguments passed to them. Removal
3383 might give a very teeny performance improvement.
3384
3385 3. Documentation bug: the value of the capture_top field in a callout is *one
3386 more than* the number of the hightest numbered captured substring.
3387
3388 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3389 in incorrectly linking with a previously installed version. They now link
3390 explicitly with libpcre.la.
3391
3392 5. configure.in no longer needs to recognize Cygwin specially.
3393
3394 6. A problem in pcre.in for Windows platforms is fixed.
3395
3396 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3397 pcretest, it used to include the size of the study block as part of its
3398 output. Unfortunately, the structure contains a field that has a different
3399 size on different hardware architectures. This meant that the tests that
3400 showed this size failed. As the block is currently always of a fixed size,
3401 this information isn't actually particularly useful in pcretest output, so
3402 I have just removed it.
3403
3404 8. Three pre-processor statements accidentally did not start in column 1.
3405 Sadly, there are *still* compilers around that complain, even though
3406 standard C has not required this for well over a decade. Sigh.
3407
3408 9. In pcretest, the code for checking callouts passed small integers in the
3409 callout_data field, which is a void * field. However, some picky compilers
3410 complained about the casts involved for this on 64-bit systems. Now
3411 pcretest passes the address of the small integer instead, which should get
3412 rid of the warnings.
3413
3414 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3415 both compile and run time, and gives an error if an invalid UTF-8 sequence
3416 is found. There is a option for disabling this check in cases where the
3417 string is known to be correct and/or the maximum performance is wanted.
3418
3419 11. In response to a bug report, I changed one line in Makefile.in from
3420
3421 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3422 to
3423 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3424
3425 to look similar to other lines, but I have no way of telling whether this
3426 is the right thing to do, as I do not use Windows. No doubt I'll get told
3427 if it's wrong...
3428
3429
3430 Version 4.3 21-May-03
3431 ---------------------
3432
3433 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3434 Makefile.
3435
3436 2. Some refactoring to improve the quality of the code:
3437
3438 (i) The utf8_table... variables are now declared "const".
3439
3440 (ii) The code for \cx, which used the "case flipping" table to upper case
3441 lower case letters, now just substracts 32. This is ASCII-specific,
3442 but the whole concept of \cx is ASCII-specific, so it seems
3443 reasonable.
3444
3445 (iii) PCRE was using its character types table to recognize decimal and
3446 hexadecimal digits in the pattern. This is silly, because it handles
3447 only 0-9, a-f, and A-F, but the character types table is locale-
3448 specific, which means strange things might happen. A private
3449 table is now used for this - though it costs 256 bytes, a table is
3450 much faster than multiple explicit tests. Of course, the standard
3451 character types table is still used for matching digits in subject
3452 strings against \d.
3453
3454 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3455 ending in _t are). So I've renamed it as ESC_tee.
3456
3457 3. The first argument for regexec() in the POSIX wrapper should have been
3458 defined as "const".
3459
3460 4. Changed pcretest to use malloc() for its buffers so that they can be
3461 Electric Fenced for debugging.
3462
3463 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3464 to read one or more bytes before the start of the subject string. Often this
3465 had no effect on PCRE's behaviour, but in some circumstances it could
3466 provoke a segmentation fault.
3467
3468 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3469 to try to read one or more bytes before the start of the subject string.
3470
3471 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3472 UTF-8 support could misbehave in various ways if the subject string
3473 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3474 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3475 back over UTF-8 characters.)
3476
3477
3478 Version 4.2 14-Apr-03
3479 ---------------------
3480
3481 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3482
3483 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3484 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3485 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3486 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3487 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3488 and BUILD_EXEEXT
3489 Note: automatic setting of the BUILD variables is not yet working
3490 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3491 compile-time but not at link-time
3492 [LINK]: use for linking executables only
3493 make different versions for Windows and non-Windows
3494 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3495 libraries
3496 [LINK_FOR_BUILD]: new variable
3497 [OBJEXT]: use throughout
3498 [EXEEXT]: use throughout
3499 <winshared>: new target
3500 <wininstall>: new target
3501 <dftables.o>: use native compiler
3502 <dftables>: use native linker
3503 <install>: handle Windows platform correctly
3504 <clean>: ditto
3505 <check>: ditto
3506 copy DLL to top builddir before testing
3507
3508 As part of these changes, -no-undefined was removed again. This was reported
3509 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3510 in any case.
3511
3512 3. Some tidies to get rid of compiler warnings:
3513
3514 . In the match_data structure, match_limit was an unsigned long int, whereas
3515 match_call_count was an int. I've made them both unsigned long ints.
3516
3517 . In pcretest the fact that a const uschar * doesn't automatically cast to
3518 a void * provoked a warning.
3519
3520 . Turning on some more compiler warnings threw up some "shadow" variables
3521 and a few more missing casts.
3522
3523 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3524 option, a class that contained a single character with a value between 128
3525 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3526
3527 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3528 option, a class that contained several characters, but with at least one
3529 whose value was between 128 and 255 caused PCRE to crash.
3530
3531
3532 Version 4.1 12-Mar-03
3533 ---------------------
3534
3535 1. Compiling with gcc -pedantic found a couple of places where casts were
3536 needed, and a string in dftables.c that was longer than standard compilers are
3537 required to support.
3538
3539 2. Compiling with Sun's compiler found a few more places where the code could
3540 be tidied up in order to avoid warnings.
3541
3542 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3543 first of these names is deprecated in the latest Autoconf in favour of the name
3544 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3545 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3546 analogy I have changed it to CFLAGS_FOR_BUILD.
3547
3548 4. Added -no-undefined to the linking command in the Makefile, because this is
3549 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3550 linking step for the pcreposix library.
3551
3552 5. PCRE was failing to diagnose the case of two named groups with the same
3553 name.
3554
3555 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3556 literal character that is needed in the subject for a match, and scans along to
3557 ensure that it is present before embarking on the full matching process. This
3558 saves time in cases of nested unlimited repeats that are never going to match.
3559 Problem: the scan can take a lot of time if the subject is very long (e.g.
3560 megabytes), thus penalizing straightforward matches. It is now done only if the
3561 amount of subject to be scanned is less than 1000 bytes.
3562
3563 7. A lesser problem with the same optimization is that it was recording the
3564 first character of an anchored pattern as "needed", thus provoking a search
3565 right along the subject, even when the first match of the pattern was going to
3566 fail. The "needed" character is now not set for anchored patterns, unless it
3567 follows something in the pattern that is of non-fixed length. Thus, it still
3568 fulfils its original purpose of finding quick non-matches in cases of nested
3569 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3570
3571
3572 Version 4.0 17-Feb-03
3573 ---------------------
3574
3575 1. If a comment in an extended regex that started immediately after a meta-item
3576 extended to the end of string, PCRE compiled incorrect data. This could lead to
3577 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3578
3579 2. Moved to autoconf 2.53 and libtool 1.4.2.
3580
3581 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3582 the special perltest8 script is no longer needed - all the tests can be run
3583 from a single perltest script.
3584
3585 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3586 by \s. It has now been removed in PCRE. This means it isn't recognized as
3587 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3588 class [:space:] *does* include VT, thereby creating a mess.
3589
3590 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3591 space and tab.
3592
3593 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3594 its new features into the main test script, reducing the number of scripts.
3595
3596 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3597 were backward compatible, and made the (?i) apply to the whole pattern, as if
3598 /i were given. Now it behaves more logically, and applies the option setting
3599 only to what follows. PCRE has been changed to follow suit. However, if it
3600 finds options settings right at the start of the pattern, it extracts them into
3601 the global options, as before. Thus, they show up in the info data.
3602
3603 8. Added support for the \Q...\E escape sequence. Characters in between are
3604 treated as literals. This is slightly different from Perl in that $ and @ are
3605 also handled as literals inside the quotes. In Perl, they will cause variable
3606 interpolation. Note the following examples:
3607
3608 Pattern PCRE matches Perl matches
3609
3610 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3611 \Qabc\$xyz\E abc\$xyz abc\$xyz
3612 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3613
3614 For compatibility with Perl, \Q...\E sequences are recognized inside character
3615 classes as well as outside them.
3616
3617 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3618 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3619 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3620 signed/unsigned warnings.
3621
3622 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3623 option for pcretest, so I've replaced it by a simple function that does just
3624 that job.
3625
3626 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3627 "pcregrep -".
3628
3629 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3630 Java package. This provides some syntactic sugar for simple cases of what my
3631 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3632 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3633 item, you can use this simplified notation. Note that only makes sense with
3634 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3635 greediness, whatever the setting of the PCRE_UNGREEDY option.
3636
3637 13. A change of greediness default within a pattern was not taking effect at
3638 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3639 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3640 was abstracted outside.
3641
3642 14. PCRE now supports the \G assertion. It is true when the current matching
3643 position is at the start point of the match. This differs from \A when the
3644 starting offset is non-zero. Used with the /g option of pcretest (or similar
3645 code), it works in the same way as it does for Perl's /g option. If all
3646 alternatives of a regex begin with \G, the expression is anchored to the start
3647 match position, and the "anchored" flag is set in the compiled expression.
3648
3649 15. Some bugs concerning the handling of certain option changes within patterns
3650 have been fixed. These applied to options other than (?ims). For example,
3651 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3652 been the other way round. Some of this was related to change 7 above.
3653
3654 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3655 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3656 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3657 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3658
3659 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3660 mode. Unlike ".", it always matches newline, whatever the setting of
3661 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3662 assertions. Perl allows it, but it doesn't (in general) work because it can't
3663 calculate the length of the lookbehind. At least, that's the case for Perl
3664 5.8.0 - I've been told they are going to document that it doesn't work in
3665 future.
3666
3667 18. Added an error diagnosis for escapes that PCRE does not support: these are
3668 \L, \l, \N, \P, \p, \U, \u, and \X.
3669
3670 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3671 reading past the end of the pattern in cases such as /[abcd/.
3672
3673 20. PCRE was getting more memory than necessary for patterns with classes that
3674 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3675
3676 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3677 compiling PCRE for use with Virtual Pascal.
3678
3679 22. Small fix to the Makefile to make it work properly if the build is done
3680 outside the source tree.
3681
3682 23. Added a new extension: a condition to go with recursion. If a conditional
3683 subpattern starts with (?(R) the "true" branch is used if recursion has
3684 happened, whereas the "false" branch is used only at the top level.
3685
3686 24. When there was a very long string of literal characters (over 255 bytes
3687 without UTF support, over 250 bytes with UTF support), the computation of how
3688 much memory was required could be incorrect, leading to segfaults or other
3689 strange effects.
3690
3691 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3692 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3693 there was a subsequent back reference to those brackets. This meant that, for
3694 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3695 possible to check for precisely this case. All we can do is abandon the
3696 optimization if .* occurs inside capturing brackets when there are any back
3697 references whatsoever. (See below for a better fix that came later.)
3698
3699 26. The handling of the optimization for finding the first character of a
3700 non-anchored pattern, and for finding a character that is required later in the
3701 match were failing in some cases. This didn't break the matching; it just
3702 failed to optimize when it could. The way this is done has been re-implemented.
3703
3704 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3705
3706 28. Added a new feature that provides some of the functionality that Perl
3707 provides with (?{...}). The facility is termed a "callout". The way it is done
3708 in PCRE is for the caller to provide an optional function, by setting
3709 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3710 global variable. By default it is unset, which disables all calling out. To get
3711 the function called, the regex must include (?C) at appropriate points. This
3712 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3713 This provides a means of identifying different callout points. When PCRE
3714 reaches such a point in the regex, if pcre_callout has been set, the external
3715 function is called. It is provided with data in a structure called
3716 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3717 matching continues; if it returns a non-zero value, the match at the current
3718 point fails. However, backtracking will occur if possible. [This was changed
3719 later and other features added - see item 49 below.]
3720
3721 29. pcretest is upgraded to test the callout functionality. It provides a
3722 callout function that displays information. By default, it shows the start of
3723 the match and the current position in the text. There are some new data escapes
3724 to vary what happens:
3725
3726 \C+ in addition, show current contents of captured substrings
3727 \C- do not supply a callout function
3728 \C!n return 1 when callout number n is reached
3729 \C!n!m return 1 when callout number n is reached for the mth time
3730
3731 30. If pcregrep was called with the -l option and just a single file name, it
3732 output "<stdin>" if a match was found, instead of the file name.
3733
3734 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3735 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3736 pcre_exec(). This saves a malloc/free per call. The default value of
3737 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3738 when configuring.
3739
3740 32. The default maximum size of a compiled pattern is 64K. There have been a
3741 few cases of people hitting this limit. The code now uses macros to handle the
3742 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3743 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3744 configuring. Tests 2 and 5 work only with 2-byte links because they output
3745 debugging information about compiled patterns.
3746
3747 33. Internal code re-arrangements:
3748
3749 (a) Moved the debugging function for printing out a compiled regex into
3750 its own source file (printint.c) and used #include to pull it into
3751 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3752 separate copies.
3753
3754 (b) Defined the list of op-code names for debugging as a macro in
3755 internal.h so that it is next to the definition of the opcodes.
3756
3757 (c) Defined a table of op-code lengths for simpler skipping along compiled
3758 code. This is again a macro in internal.h so that it is next to the
3759 definition of the opcodes.
3760
3761 34. Added support for recursive calls to individual subpatterns, along the
3762 lines of Robin Houston's patch (but implemented somewhat differently).
3763
3764 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3765 allow it to read and process whole directories in Win32. This code was
3766 contributed by Lionel Fourquaux; it has not been tested by me.
3767
3768 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3769 used to name a group. Names consist of alphanumerics and underscores, and must
3770 be unique. Back references use the syntax (?P=name) and recursive calls use
3771 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3772 numbers. The function pcre_fullinfo() can be used after compilation to extract
3773 a name/number map. There are three relevant calls:
3774
3775 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3776 PCRE_INFO_NAMECOUNT yields the number of entries
3777 PCRE_INFO_NAMETABLE yields a pointer to the map.
3778
3779 The map is a vector of fixed-size entries. The size of each entry depends on
3780 the length of the longest name used. The first two bytes of each entry are the
3781 group number, most significant byte first. There follows the corresponding
3782 name, zero terminated. The names are in alphabetical order.
3783
3784 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3785 case instead of 255. Making it the same both with and without UTF-8 support
3786 means that the same test output works with both.
3787
3788 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3789 calling malloc() with a zero argument.
3790
3791 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3792 optimization. I've improved things by keeping a bitmap of backreferences with
3793 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3794 fact referenced, the optimization can be applied. It is unlikely that a
3795 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3796 the match to follow \n) will appear inside brackets with a number greater than
3797 31, but if it does, any back reference > 31 suppresses the optimization.
3798
3799 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3800 of disabling numbered capturing parentheses. Any opening parenthesis that is
3801 not followed by ? behaves as if it were followed by ?: but named parentheses
3802 can still be used for capturing (and they will acquire numbers in the usual
3803 way).
3804
3805 41. Redesigned the return codes from the match() function into yes/no/error so
3806 that errors can be passed back from deep inside the nested calls. A malloc
3807 failure while inside a recursive subpattern call now causes the
3808 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3809
3810 42. It is now possible to set a limit on the number of times the match()
3811 function is called in a call to pcre_exec(). This facility makes it possible to
3812 limit the amount of recursion and backtracking, though not in a directly
3813 obvious way, because the match() function is used in a number of different
3814 circumstances. The count starts from zero for each position in the subject
3815 string (for non-anchored patterns). The default limit is, for compatibility, a
3816 large number, namely 10 000 000. You can change this in two ways:
3817
3818 (a) When configuring PCRE before making, you can use --with-match-limit=n
3819 to set a default value for the compiled library.
3820
3821 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3822 a different value is set. See 45 below.
3823
3824 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3825
3826 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3827 of things that can be changed at compile time. The first argument specifies
3828 what is wanted and the second points to where the information is to be placed.
3829 The current list of available information is:
3830
3831 PCRE_CONFIG_UTF8
3832
3833 The output is an integer that is set to one if UTF-8 support is available;
3834 otherwise it is set to zero.
3835
3836 PCRE_CONFIG_NEWLINE
3837
3838 The output is an integer that it set to the value of the code that is used for
3839 newline. It is either LF (10) or CR (13).
3840
3841 PCRE_CONFIG_LINK_SIZE
3842
3843 The output is an integer that contains the number of bytes used for internal
3844 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3845
3846 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3847
3848 The output is an integer that contains the threshold above which the POSIX
3849 interface uses malloc() for output vectors. See item 31 above.
3850
3851 PCRE_CONFIG_MATCH_LIMIT
3852
3853 The output is an unsigned integer that contains the default limit of the number
3854 of match() calls in a pcre_exec() execution. See 42 above.
3855
3856 44. pcretest has been upgraded by the addition of the -C option. This causes it
3857 to extract all the available output from the new pcre_config() function, and to
3858 output it. The program then exits immediately.
3859
3860 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3861 order to support additional features. One way would have been to define
3862 pcre_exec2() (for example) with extra arguments, but this would not have been
3863 extensible, and would also have required all calls to the original function to
3864 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3865 is used for passing in "extra" data from pcre_study().
3866
3867 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3868 contains the following fields:
3869
3870 flags a bitmap indicating which of the following fields are set
3871 study_data opaque data from pcre_study()
3872 match_limit a way of specifying a limit on match() calls for a specific
3873 call to pcre_exec()
3874 callout_data data for callouts (see 49 below)
3875
3876 The flag bits are also defined in pcre.h, and are
3877
3878 PCRE_EXTRA_STUDY_DATA
3879 PCRE_EXTRA_MATCH_LIMIT
3880 PCRE_EXTRA_CALLOUT_DATA
3881
3882 The pcre_study() function now returns one of these new pcre_extra blocks, with
3883 the actual study data pointed to by the study_data field, and the
3884 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3885 before. That is, this change is entirely upwards-compatible and requires no
3886 change to existing code.
3887
3888 If you want to pass in additional data to pcre_exec(), you can either place it
3889 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3890 block.
3891
3892 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3893 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3894 times with different match limits, until it finds the minimum value needed for
3895 pcre_exec() to complete. The value is then output. This can be instructive; for
3896 most simple matches the number is quite small, but for pathological cases it
3897 gets very large very quickly.
3898
3899 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3900 returns the size of the data block pointed to by the study_data field in a
3901 pcre_extra block, that is, the value that was passed as the argument to
3902 pcre_malloc() when PCRE was getting memory in which to place the information
3903 created by pcre_study(). The fourth argument should point to a size_t variable.
3904 pcretest has been extended so that this information is shown after a successful
3905 pcre_study() call when information about the compiled regex is being displayed.
3906
3907 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3908 because what follows is always an absolute path. (Later: it turns out that this
3909 is more than cosmetic for MinGW, because it doesn't like empty path
3910 components.)
3911
3912 49. Some changes have been made to the callout feature (see 28 above):
3913
3914 (i) A callout function now has three choices for what it returns:
3915
3916 0 => success, carry on matching
3917 > 0 => failure at this point, but backtrack if possible
3918 < 0 => serious error, return this value from pcre_exec()
3919
3920 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3921 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3922 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3923 use by callout functions. It will never be used by PCRE itself.
3924
3925 (ii) The pcre_extra structure (see 45 above) has a void * field called
3926 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3927 pcre_callout_block structure has a field of the same name. The contents of
3928 the field passed in the pcre_extra structure are passed to the callout
3929 function in the corresponding field in the callout block. This makes it
3930 easier to use the same callout-containing regex from multiple threads. For
3931 testing, the pcretest program has a new data escape
3932
3933 \C*n pass the number n (may be negative) as callout_data
3934
3935 If the callout function in pcretest receives a non-zero value as
3936 callout_data, it returns that value.
3937
3938 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3939 there were some redundant $(CFLAGS) in commands that are now specified as
3940 $(LINK), which already includes $(CFLAGS).
3941
3942 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3943 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3944 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3945 one-byte characters throughout. Note that case-insensitive matching applies
3946 only to characters whose values are less than 256. PCRE doesn't support the
3947 notion of cases for higher-valued characters.
3948
3949 (i) A character class whose characters are all within 0-255 is handled as
3950 a bit map, and the map is inverted for negative classes. Previously, a
3951 character > 255 always failed to match such a class; however it should
3952 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3953
3954 (ii) A negated character class with a single character < 255 is coded as
3955 "not this character" (OP_NOT). This wasn't working properly when the test
3956 character was multibyte, either singly or repeated.
3957
3958 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3959 mode, for example: \x{100}{2,3}.
3960
3961 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3962 singly or repeated) now correctly test multibyte characters. However,
3963 PCRE doesn't recognize any characters with values greater than 255 as
3964 digits, spaces, or word characters. Such characters always match \D, \S,
3965 and \W, and never match \d, \s, or \w.
3966
3967 (v) Classes may now contain characters and character ranges with values
3968 greater than 255. For example: [ab\x{100}-\x{400}].
3969
3970 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3971 PCRE in UTF-8 mode.
3972
3973 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3974 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3975 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3976 value.)
3977
3978 53. The single man page has become too large. I have therefore split it up into
3979 a number of separate man pages. These also give rise to individual HTML pages;
3980 these are now put in a separate directory, and there is an index.html page that
3981 lists them all. Some hyperlinking between the pages has been installed.
3982
3983 54. Added convenience functions for handling named capturing parentheses.
3984
3985 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3986 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3987 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3988 are faulted.
3989
3990 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3991 calling configure. These values are used when compiling the dftables.c program
3992 which is run to generate the source of the default character tables. They
3993 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3994 you will need to set these values.
3995
3996 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3997
3998
3999 Version 3.9 02-Jan-02
4000 ---------------------
4001
4002 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4003
4004 2. If --disable-static was given, the building process failed when trying to
4005 build pcretest and pcregrep. (For some reason it was using libtool to compile
4006 them, which is not right, as they aren't part of the library.)
4007
4008
4009 Version 3.8 18-Dec-01
4010 ---------------------
4011
4012 1. The experimental UTF-8 code was completely screwed up. It was packing the
4013 bytes in the wrong order. How dumb can you get?
4014
4015
4016 Version 3.7 29-Oct-01
4017 ---------------------
4018
4019 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4020 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4021 this didn't happen under Solaris 8, where I normally test things.
4022
4023 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4024 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4025 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4026 this fix an hour or so after the initial 3.7 release.)
4027
4028
4029 Version 3.6 23-Oct-01
4030 ---------------------
4031
4032 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4033 offsets passed as NULL with zero offset count.
4034
4035 2. The config.guess and config.sub files had not been updated when I moved to
4036 the latest autoconf.
4037
4038
4039 Version 3.5 15-Aug-01
4040 ---------------------
4041
4042 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4043 had been forgotten.
4044
4045 2. By using declared but undefined structures, we can avoid using "void"
4046 definitions in pcre.h while keeping the internal definitions of the structures
4047 private.
4048
4049 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4050 user point of view, this means that both static and shared libraries are built
4051 by default, but this can be individually controlled. More of the work of
4052 handling this static/shared cases is now inside libtool instead of PCRE's make
4053 file.
4054
4055 4. The pcretest utility is now installed along with pcregrep because it is
4056 useful for users (to test regexs) and by doing this, it automatically gets
4057 relinked by libtool. The documentation has been turned into a man page, so
4058 there are now .1, .txt, and .html versions in /doc.
4059
4060 5. Upgrades to pcregrep:
4061 (i) Added long-form option names like gnu grep.
4062 (ii) Added --help to list all options with an explanatory phrase.
4063 (iii) Added -r, --recursive to recurse into sub-directories.
4064 (iv) Added -f, --file to read patterns from a file.
4065
4066 6. pcre_exec() was referring to its "code" argument before testing that
4067 argument for NULL (and giving an error if it was NULL).
4068
4069 7. Upgraded Makefile.in to allow for compiling in a different directory from
4070 the source directory.
4071
4072 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4073 options bits, the pointer it was passed was to an int instead of to an unsigned
4074 long int. This mattered only on 64-bit systems.
4075
4076 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4077 generated) instead of pcre.in, which it its source. Also made the same change
4078 in several of the .c files.
4079
4080 10. A new release of gcc defines printf() as a macro, which broke pcretest
4081 because it had an ifdef in the middle of a string argument for printf(). Fixed
4082 by using separate calls to printf().
4083
4084 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4085 script, to force use of CR or LF instead of \n in the source. On non-Unix
4086 systems, the value can be set in config.h.
4087
4088 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4089 absolute limit. Changed the text of the error message to make this clear, and
4090 likewise updated the man page.
4091
4092 13. The limit of 99 on the number of capturing subpatterns has been removed.
4093 The new limit is 65535, which I hope will not be a "real" limit.
4094
4095
4096 Version 3.4 22-Aug-00
4097 ---------------------
4098
4099 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4100
4101 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4102
4103
4104 Version 3.3 01-Aug-00
4105 ---------------------
4106
4107 1. If an octal character was given, but the value was greater than \377, it
4108 was not getting masked to the least significant bits, as documented. This could
4109 lead to crashes in some systems.
4110
4111 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4112 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4113
4114 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4115 These just pass their arguments on to (pcre_free)(), but they are provided
4116 because some uses of PCRE bind it to non-C systems that can call its functions,
4117 but cannot call free() or pcre_free() directly.
4118
4119 4. Add "make test" as a synonym for "make check". Corrected some comments in
4120 the Makefile.
4121
4122 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4123 Makefile.
4124
4125 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4126 command called pgrep for grepping around the active processes.
4127
4128 7. Added the beginnings of support for UTF-8 character strings.
4129
4130 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4131 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4132 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4133 out for the ar command.)
4134
4135
4136 Version 3.2 12-May-00
4137 ---------------------
4138
4139 This is purely a bug fixing release.
4140
4141 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4142 of ZA. This was just one example of several cases that could provoke this bug,
4143 which was introduced by change 9 of version 2.00. The code for breaking
4144 infinite loops after an iteration that matches an empty string was't working
4145 correctly.
4146
4147 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4148 when matched against abbab (for example). After matching an empty string, it
4149 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4150 caused it to match further down the string than it should.
4151
4152 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4153 was there because it doesn't seem to be needed, and it causes trouble on some
4154 systems, as it is not a Standard C header. It has been removed.
4155
4156 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4157 were reported on the Macintosh. The changes were from
4158
4159 while ((c = *(++ptr)) != 0 && c != '\n');
4160 to
4161 while ((c = *(++ptr)) != 0 && c != '\n') ;
4162
4163 Totally extraordinary, but if that's what it takes...
4164
4165 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4166 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4167 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4168 assumes the way PCRE uses memmove() (always moving upwards).
4169
4170 6. PCRE is being used in one environment where strchr() is not available. There
4171 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4172 faster code anyway.
4173
4174
4175 Version 3.1 09-Feb-00
4176 ---------------------
4177
4178 The only change in this release is the fixing of some bugs in Makefile.in for
4179 the "install" target:
4180
4181 (1) It was failing to install pcreposix.h.
4182
4183 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4184
4185
4186 Version 3.0 01-Feb-00
4187 ---------------------
4188
4189 1. Add support for the /+ modifier to perltest (to output $` like it does in
4190 pcretest).
4191
4192 2. Add support for the /g modifier to perltest.
4193
4194 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4195 matches null strings.
4196
4197 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4198 pattern. Perl treats empty patterns specially - it reuses the most recent
4199 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4200 effect.
4201
4202 5. The POSIX interface was broken in that it was just handing over the POSIX
4203 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4204 required a bigger vector, with some working space on the end. This means that
4205 the POSIX wrapper now has to get and free some memory, and copy the results.
4206
4207 6. Added some simple autoconf support, placing the test data and the
4208 documentation in separate directories, re-organizing some of the
4209 information files, and making it build pcre-config (a GNU standard). Also added
4210 libtool support for building PCRE as a shared library, which is now the
4211 default.
4212
4213 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4214 09 are not valid octal constants. Single digits will be used for minor values
4215 less than 10.
4216
4217 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4218 existing programs that set these in the POSIX interface can use PCRE without
4219 modification.
4220
4221 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4222 return all that pcre_info() returns, plus additional data. The pcre_info()
4223 function is retained for compatibility, but is considered to be obsolete.
4224
4225 10. Added experimental recursion feature (?R) to handle one common case that
4226 Perl 5.6 will be able to do with (?p{...}).
4227
4228 11. Added support for POSIX character classes like [:alpha:], which Perl is
4229 adopting.
4230
4231
4232 Version 2.08 31-Aug-99
4233 ----------------------
4234
4235 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4236 trying to match at the startoffset position, but instead was moving forward to
4237 the next newline as if a previous match had failed.
4238
4239 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4240 and could get into a loop if a null string was matched other than at the start
4241 of the subject.
4242
4243 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4244 be distinguished at compile time, and for completeness also added PCRE_DATE.
4245
4246 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4247 in GnuWin32 environments.
4248
4249
4250 Version 2.07 29-Jul-99
4251 ----------------------
4252
4253 1. The documentation is now supplied in plain text form and HTML as well as in
4254 the form of man page sources.
4255
4256 2. C++ compilers don't like assigning (void *) values to other pointer types.
4257 In particular this affects malloc(). Although there is no problem in Standard
4258 C, I've put in casts to keep C++ compilers happy.
4259
4260 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4261 should be (const char *).
4262
4263 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4264 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4265 However, I haven't made this a standard facility. The documentation doesn't
4266 mention it, and the Makefile doesn't support it.
4267
4268 5. The Makefile now contains an "install" target, with editable destinations at
4269 the top of the file. The pcretest program is not installed.
4270
4271 6. pgrep -V now gives the PCRE version number and date.
4272
4273 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4274 causing the entire string to be ignored, instead of just the last character.
4275
4276 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4277 non-matching string, it can take a very, very long time, even for strings of
4278 quite modest length, because of the nested recursion. PCRE now does better in
4279 some of these cases. It does this by remembering the last required literal
4280 character in the pattern, and pre-searching the subject to ensure it is present
4281 before running the real match. In other words, it applies a heuristic to detect
4282 some types of certain failure quickly, and in the above example, if presented
4283 with a string that has no trailing " it gives "no match" very quickly.
4284
4285 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4286 other alternatives are tried instead.
4287
4288
4289 Version 2.06 09-Jun-99
4290 ----------------------
4291
4292 1. Change pcretest's output for amount of store used to show just the code
4293 space, because the remainder (the data block) varies in size between 32-bit and
4294 64-bit systems.
4295
4296 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4297 start matching at. This allows lookbehinds to work when searching for multiple
4298 occurrences in a string.
4299
4300 3. Added additional options to pcretest for testing multiple occurrences:
4301
4302 /+ outputs the rest of the string that follows a match
4303 /g loops for multiple occurrences, using the new startoffset argument
4304 /G loops for multiple occurrences by passing an incremented pointer
4305
4306 4. PCRE wasn't doing the "first character" optimization for patterns starting
4307 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4308 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4309 the letter 'x'. On long subject strings, this gives a significant speed-up.
4310
4311
4312 Version 2.05 21-Apr-99
4313 ----------------------
4314
4315 1. Changed the type of magic_number from int to long int so that it works
4316 properly on 16-bit systems.
4317
4318 2. Fixed a bug which caused patterns starting with .* not to work correctly
4319 when the subject string contained newline characters. PCRE was assuming
4320 anchoring for such patterns in all cases, which is not correct because .* will
4321 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4322 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4323 must be retried after every newline in the subject.
4324
4325
4326 Version 2.04 18-Feb-99
4327 ----------------------
4328
4329 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4330 computation of the store needed to hold the pattern was incorrect (too large).
4331 If such patterns were nested a few deep, this could multiply and become a real
4332 problem.
4333
4334 2. Added /M option to pcretest to show the memory requirement of a specific
4335 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4336
4337 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4338 compiled in such a way that the backtracking after subsequent failure was
4339 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4340 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4341
4342
4343 Version 2.03 02-Feb-99
4344 ----------------------
4345
4346 1. Fixed typo and small mistake in man page.
4347
4348 2. Added 4th condition (GPL supersedes if conflict) and created separate
4349 LICENCE file containing the conditions.
4350
4351 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4352 Perl, that is the internal \ allows the delimiter to be included in the
4353 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4354 the final delimiter, add \ to the end of the pattern (to test the error).
4355
4356 4. Added the convenience functions for extracting substrings after a successful
4357 match. Updated pcretest to make it able to test these functions.
4358
4359
4360 Version 2.02 14-Jan-99
4361 ----------------------
4362
4363 1. Initialized the working variables associated with each extraction so that
4364 their saving and restoring doesn't refer to uninitialized store.
4365
4366 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4367 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4368 fix the problem.
4369
4370 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4371 calls, and wasn't printing the correct value for compiling calls. Increased the
4372 default value of LOOPREPEAT, and the number of significant figures in the
4373 times.
4374
4375 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4376
4377 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4378 a building problem on Windows NT with a FAT file system.
4379
4380
4381 Version 2.01 21-Oct-98
4382 ----------------------
4383
4384 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4385 to character tables built by pcre_maketables() in the current locale. If NULL
4386 is passed, the default tables are used.
4387
4388
4389 Version 2.00 24-Sep-98
4390 ----------------------
4391
4392 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4393 it any more.
4394
4395 2. Allow quantification of (?>) groups, and make it work correctly.
4396
4397 3. The first character computation wasn't working for (?>) groups.
4398
4399 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4400 end of the subject) and add 5.005's \z, which really does match only at the
4401 very end of the subject.
4402
4403 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4404
4405 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4406 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4407 localized options. All options to pcre_study() were also removed.
4408
4409 7. Add other new features from 5.005:
4410
4411 $(?<= positive lookbehind
4412 $(?<! negative lookbehind
4413 (?imsx-imsx) added the unsetting capability
4414 such a setting is global if at outer level; local otherwise
4415 (?imsx-imsx:) non-capturing groups with option setting
4416 (?(cond)re|re) conditional pattern matching
4417
4418 A backreference to itself in a repeated group matches the previous
4419 captured string.
4420
4421 8. General tidying up of studying (both automatic and via "study")
4422 consequential on the addition of new assertions.
4423
4424 9. As in 5.005, unlimited repeated groups that could match an empty substring
4425 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4426 runtime if any iteration does actually match an empty substring.
4427
4428 10. Include the RunTest script in the distribution.
4429
4430 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4431 discrepancies, some of which were old and were also with respect to 5.004. They
4432 have now been fixed.
4433
4434
4435 Version 1.09 28-Apr-98
4436 ----------------------
4437
4438 1. A negated single character class followed by a quantifier with a minimum
4439 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4440 program crashes, or just wrong answers. This did not apply to negated classes
4441 containing more than one character, or to minima other than one.
4442
4443
4444 Version 1.08 27-Mar-98
4445 ----------------------
4446
4447 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4448
4449 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4450 latter must appear before anything that relies on it in the pattern.
4451
4452
4453 Version 1.07 16-Feb-98
4454 ----------------------
4455
4456 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4457 repeat of a potentially empty string).
4458
4459
4460 Version 1.06 23-Jan-98
4461 ----------------------
4462
4463 1. Added Markus Oberhumer's little patches for C++.
4464
4465 2. Literal strings longer than 255 characters were broken.
4466
4467
4468 Version 1.05 23-Dec-97
4469 ----------------------
4470
4471 1. Negated character classes containing more than one character were failing if
4472 PCRE_CASELESS was set at run time.
4473
4474
4475 Version 1.04 19-Dec-97
4476 ----------------------
4477
4478 1. Corrected the man page, where some "const" qualifiers had been omitted.
4479
4480 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4481 input syntax.
4482
4483 3. Fixed memory leak which occurred when a regex with back references was
4484 matched with an offsets vector that wasn't big enough. The temporary memory
4485 that is used in this case wasn't being freed if the match failed.
4486
4487 4. Tidied pcretest to ensure it frees memory that it gets.
4488
4489 5. Temporary memory was being obtained in the case where the passed offsets
4490 vector was exactly big enough.
4491
4492 6. Corrected definition of offsetof() from change 5 below.
4493
4494 7. I had screwed up change 6 below and broken the rules for the use of
4495 setjmp(). Now fixed.
4496
4497
4498 Version 1.03 18-Dec-97
4499 ----------------------
4500
4501 1. A erroneous regex with a missing opening parenthesis was correctly
4502 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4503 on some systems.
4504
4505 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4506 it was reported that one broken compiler failed on the former because "code" is
4507 also an independent variable.
4508
4509 3. The erroneous regex a[]b caused an array overrun reference.
4510
4511 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4512 fail on data ending with that character. (It was going on too far, and checking
4513 the next character, typically a binary zero.) This was specific to the
4514 optimized code for single-character negative classes.
4515
4516 5. Added a contributed patch from the TIN world which does the following:
4517
4518 + Add an undef for memmove, in case the the system defines a macro for it.
4519
4520 + Add a definition of offsetof(), in case there isn't one. (I don't know
4521 the reason behind this - offsetof() is part of the ANSI standard - but
4522 it does no harm).
4523
4524 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4525 most of the places where whitespace preceded '#'. I have given up and
4526 allowed the remaining 2 cases to be at the margin.
4527
4528 + Rename some variables in pcre to eliminate shadowing. This seems very
4529 pedantic, but does no harm, of course.
4530
4531 6. Moved the call to setjmp() into its own function, to get rid of warnings
4532 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4533
4534 7. Constructs such as \d{8,} were compiling into the equivalent of
4535 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4536 outcome, but in this particular case used more store than had been allocated,
4537 which caused the bug to be discovered because it threw up an internal error.
4538
4539 8. The debugging code in both pcre and pcretest for outputting the compiled
4540 form of a regex was going wrong in the case of back references followed by
4541 curly-bracketed repeats.
4542
4543
4544 Version 1.02 12-Dec-97
4545 ----------------------
4546
4547 1. Typos in pcre.3 and comments in the source fixed.
4548
4549 2. Applied a contributed patch to get rid of places where it used to remove
4550 'const' from variables, and fixed some signed/unsigned and uninitialized
4551 variable warnings.
4552
4553 3. Added the "runtest" target to Makefile.
4554
4555 4. Set default compiler flag to -O2 rather than just -O.
4556
4557
4558 Version 1.01 19-Nov-97
4559 ----------------------
4560
4561 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4562 like /([ab]*)*/, that is, for classes with more than one character in them.
4563
4564 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4565 as /((?>a*))*/ (a PCRE_EXTRA facility).
4566
4567
4568 Version 1.00 18-Nov-97
4569 ----------------------
4570
4571 1. Added compile-time macros to support systems such as SunOS4 which don't have
4572 memmove() or strerror() but have other things that can be used instead.
4573
4574 2. Arranged that "make clean" removes the executables.
4575
4576
4577 Version 0.99 27-Oct-97
4578 ----------------------
4579
4580 1. Fixed bug in code for optimizing classes with only one character. It was
4581 initializing a 32-byte map regardless, which could cause it to run off the end
4582 of the memory it had got.
4583
4584 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4585
4586
4587 Version 0.98 22-Oct-97
4588 ----------------------
4589
4590 1. Fixed bug in code for handling temporary memory usage when there are more
4591 back references than supplied space in the ovector. This could cause segfaults.
4592
4593
4594 Version 0.97 21-Oct-97
4595 ----------------------
4596
4597 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4598
4599 2. Optimized negated single characters not to use a bit map.
4600
4601 3. Brought error texts together as macro definitions; clarified some of them;
4602 fixed one that was wrong - it said "range out of order" when it meant "invalid
4603 escape sequence".
4604
4605 4. Changed some char * arguments to const char *.
4606
4607 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4608
4609 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4610 pcretest.
4611
4612
4613 Version 0.96 16-Oct-97
4614 ----------------------
4615
4616 1. Added a simple "pgrep" utility to the distribution.
4617
4618 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4619 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4620 where "ddd" means "one or more decimal digits".
4621
4622 3. Fixed serious bug. If a pattern had a back reference, but the call to
4623 pcre_exec() didn't supply a large enough ovector to record the related
4624 identifying subpattern, the match always failed. PCRE now remembers the number
4625 of the largest back reference, and gets some temporary memory in which to save
4626 the offsets during matching if necessary, in order to ensure that
4627 backreferences always work.
4628
4629 4. Increased the compatibility with Perl in a number of ways:
4630
4631 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4632 to request this handling. The option can be set at compile or exec time.
4633
4634 (b) $ matches before a terminating newline by default; an option
4635 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4636 mode). The option can be set at compile or exec time.
4637
4638 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4639 the same as Perl's. If the decimal number it represents is less than 10
4640 or there aren't that many previous left capturing parentheses, an octal
4641 escape is read. Inside a character class, it's always an octal escape,
4642 even if it is a single digit.
4643
4644 (d) An escaped but undefined alphabetic character is taken as a literal,
4645 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4646 escapes.
4647
4648 (e) {0} is now permitted. (The previous item is removed from the compiled
4649 pattern).
4650
4651 5. Changed all the names of code files so that the basic parts are no longer
4652 than 10 characters, and abolished the teeny "globals.c" file.
4653
4654 6. Changed the handling of character classes; they are now done with a 32-byte
4655 bit map always.
4656
4657 7. Added the -d and /D options to pcretest to make it possible to look at the
4658 internals of compilation without having to recompile pcre.
4659
4660
4661 Version 0.95 23-Sep-97
4662 ----------------------
4663
4664 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4665 \x20 at the start of a run of normal characters. These were being treated as
4666 real characters, instead of the source characters being re-checked.
4667
4668
4669 Version 0.94 18-Sep-97
4670 ----------------------
4671
4672 1. The functions are now thread-safe, with the caveat that the global variables
4673 containing pointers to malloc() and free() or alternative functions are the
4674 same for all threads.
4675
4676 2. Get pcre_study() to generate a bitmap of initial characters for non-
4677 anchored patterns when this is possible, and use it if passed to pcre_exec().
4678
4679
4680 Version 0.93 15-Sep-97
4681 ----------------------
4682
4683 1. /(b)|(:+)/ was computing an incorrect first character.
4684
4685 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4686 but not actually doing anything yet.
4687
4688 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4689 as Perl does (e.g. [-az] or [az-]).
4690
4691 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4692 all possible positions.
4693
4694 5. Split up into different modules to avoid including unneeded functions in a
4695 compiled binary. However, compile and exec are still in one module. The "study"
4696 function is split off.
4697
4698 6. The character tables are now in a separate module whose source is generated
4699 by an auxiliary program - but can then be edited by hand if required. There are
4700 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4701 toupper() in the code.
4702
4703 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4704 make them global. Abolish the function for setting them, as the caller can now
4705 set them directly.
4706
4707
4708 Version 0.92 11-Sep-97
4709 ----------------------
4710
4711 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4712 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4713
4714 2. Caseless matching was not working in character classes if the characters in
4715 the pattern were in upper case.
4716
4717 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4718
4719 4. Make PCRE_ANCHORED public and accept as a compile option.
4720
4721 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4722 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4723 pass them.
4724
4725 6. Give an error if bad option bits passed at compile or run time.
4726
4727 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4728 pcretest to cause it to pass that flag.
4729
4730 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4731 options, and the first character, if set.
4732
4733 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4734
4735
4736 Version 0.91 10-Sep-97
4737 ----------------------
4738
4739 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4740 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4741
4742 2. PCRE was looping on encountering an indefinitely repeated back reference to
4743 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4744 Perl does - treats the match as successful.
4745
4746 ****

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12