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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 934 - (show annotations) (download)
Sat Feb 25 12:30:36 2012 UTC (2 years, 9 months ago) by ph10
File size: 228225 byte(s)
Stop (*COMMIT) escaping from a recursive subroutine call.

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

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