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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 924 - (show annotations) (download)
Wed Feb 22 10:23:56 2012 UTC (2 years, 8 months ago) by zherczeg
File size: 227706 byte(s)
OP_NOT now supports any UTF character
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. When matching is performed using JIT-compiled code, the flag bit
37 PCRE_EXTRA_USED_JIT is set in the pcre_extra block and can be tested by the
38 caller. This bit is checked by pcretest if JIT is requested by /S++ or -s++
39 (instead of just /S+ or -s+) and the text "(JIT)" added to the output if
40 the bit is set.
41
42 7. Individual JIT compile options can be set in pcretest by following -s+[+]
43 or /S+[+] with a digit between 1 and 7.
44
45 8. OP_NOT now supports any UTF character not just single character ones.
46
47
48 Version 8.30 04-February-2012
49 -----------------------------
50
51 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
52 name is defined in ctype.h.
53
54 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
55 only in quite long subpatterns.
56
57 3. Removed the function pcre_info(), which has been obsolete and deprecated
58 since it was replaced by pcre_fullinfo() in February 2000.
59
60 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
61 match a (*MARK), and the match failed at the start of the subject, a
62 reference to memory before the start of the subject could occur. This bug
63 was introduced by fix 17 of release 8.21.
64
65 5. A reference to an unset group with zero minimum repetition was giving
66 totally wrong answers (in non-JavaScript-compatibility mode). For example,
67 /(another)?(\1?)test/ matched against "hello world test". This bug was
68 introduced in release 8.13.
69
70 6. Add support for 16-bit character strings (a large amount of work involving
71 many changes and refactorings).
72
73 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
74 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
75 from a file.
76
77 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
78 rounding is not applied in this particular case).
79
80 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
81 if they appear, or are escaped, in patterns.
82
83 10. Get rid of a number of -Wunused-but-set-variable warnings.
84
85 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
86 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
87 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
88 also returns the mark "x". This bug applied to capturing parentheses,
89 non-capturing parentheses, and atomic parentheses. It also applied to some
90 assertions.
91
92 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
93 information out of configure.ac instead of relying on pcre.h.generic, which
94 is not stored in the repository.
95
96 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
97 -lreadline.
98
99 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
100
101 15. Applied Graycode's patch to put the top-level frame on the stack rather
102 than the heap when not using the stack for recursion. This gives a
103 performance improvement in many cases when recursion is not deep.
104
105 16. Experimental code added to "pcretest -C" to output the stack frame size.
106
107
108 Version 8.21 12-Dec-2011
109 ------------------------
110
111 1. Updating the JIT compiler.
112
113 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
114 are added as well.
115
116 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
117 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
118 calling _pcre_jit_exec. Some extra comments are added.
119
120 4. (*MARK) settings inside atomic groups that do not contain any capturing
121 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
122 was introduced by change 18 for 8.20.
123
124 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
125 ECMA-262 standard.
126
127 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
128 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
129 This bug was probably introduced by change 9 of 8.13.
130
131 7. While fixing 6 above, I noticed that a number of other items were being
132 incorrectly rejected as "not fixed length". This arose partly because newer
133 opcodes had not been added to the fixed-length checking code. I have (a)
134 corrected the bug and added tests for these items, and (b) arranged for an
135 error to occur if an unknown opcode is encountered while checking for fixed
136 length instead of just assuming "not fixed length". The items that were
137 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
138 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
139 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
140
141 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
142 being incorrectly compiled and would have given unpredicatble results.
143
144 9. A possessively repeated subpattern with minimum repeat count greater than
145 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
146 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
147 the first (A) could occur when it should not.
148
149 10. Add a cast and remove a redundant test from the code.
150
151 11. JIT should use pcre_malloc/pcre_free for allocation.
152
153 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
154 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
155 is anything other than /usr, -L is still shown).
156
157 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
158
159 14. Perl does not support \N without a following name in a [] class; PCRE now
160 also gives an error.
161
162 15. If a forward reference was repeated with an upper limit of around 2000,
163 it caused the error "internal error: overran compiling workspace". The
164 maximum number of forward references (including repeats) was limited by the
165 internal workspace, and dependent on the LINK_SIZE. The code has been
166 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
167 the default depends on LINK_SIZE. There is a new upper limit (for safety)
168 of around 200,000 forward references. While doing this, I also speeded up
169 the filling in of repeated forward references.
170
171 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
172 incorrectly expecting the subject to contain another "a" after the start.
173
174 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
175 in the match, the SKIP should be ignored. This was not happening; instead
176 the SKIP was being treated as NOMATCH. For patterns such as
177 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
178 tested.
179
180 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
181 now much more compatible with Perl, in particular in cases where the result
182 is a non-match for a non-anchored pattern. For example, if
183 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
184 "m", where previously it did not return a name. A side effect of this
185 change is that for partial matches, the last encountered mark name is
186 returned, as for non matches. A number of tests that were previously not
187 Perl-compatible have been moved into the Perl-compatible test files. The
188 refactoring has had the pleasing side effect of removing one argument from
189 the match() function, thus reducing its stack requirements.
190
191 19. If the /S+ option was used in pcretest to study a pattern using JIT,
192 subsequent uses of /S (without +) incorrectly behaved like /S+.
193
194 21. Retrieve executable code size support for the JIT compiler and fixing
195 some warnings.
196
197 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
198 not work when the shorter character appeared right at the end of the
199 subject string.
200
201 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
202 systems.
203
204 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
205 output it when the /M option is used in pcretest.
206
207 25. The CheckMan script was not being included in the distribution. Also, added
208 an explicit "perl" to run Perl scripts from the PrepareRelease script
209 because this is reportedly needed in Windows.
210
211 26. If study data was being save in a file and studying had not found a set of
212 "starts with" bytes for the pattern, the data written to the file (though
213 never used) was taken from uninitialized memory and so caused valgrind to
214 complain.
215
216 27. Updated RunTest.bat as provided by Sheri Pierce.
217
218 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
219
220 29. Computation of memory usage for the table of capturing group names was
221 giving an unnecessarily large value.
222
223
224 Version 8.20 21-Oct-2011
225 ------------------------
226
227 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
228 a POSIX class. After further experiments with Perl, which convinced me that
229 Perl has bugs and confusions, a closing square bracket is no longer allowed
230 in a POSIX name. This bug also affected patterns with classes that started
231 with full stops.
232
233 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
234 captured substring, but while checking the failing first alternative,
235 substring 1 is temporarily captured. If the output vector supplied to
236 pcre_exec() was not big enough for this capture, the yield of the function
237 was still zero ("insufficient space for captured substrings"). This cannot
238 be totally fixed without adding another stack variable, which seems a lot
239 of expense for a edge case. However, I have improved the situation in cases
240 such as /(a)(b)x|abc/ matched against "abc", where the return code
241 indicates that fewer than the maximum number of slots in the ovector have
242 been set.
243
244 3. Related to (2) above: when there are more back references in a pattern than
245 slots in the output vector, pcre_exec() uses temporary memory during
246 matching, and copies in the captures as far as possible afterwards. It was
247 using the entire output vector, but this conflicts with the specification
248 that only 2/3 is used for passing back captured substrings. Now it uses
249 only the first 2/3, for compatibility. This is, of course, another edge
250 case.
251
252 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
253 main code base, and can be used by building with --enable-jit. When this is
254 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
255 runtime --no-jit option is given.
256
257 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
258 ovector, the return from the function was zero, implying that there were
259 other matches that did not fit. The correct "exactly full" value is now
260 returned.
261
262 6. If a subpattern that was called recursively or as a subroutine contained
263 (*PRUNE) or any other control that caused it to give a non-standard return,
264 invalid errors such as "Error -26 (nested recursion at the same subject
265 position)" or even infinite loops could occur.
266
267 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
268 computing the minimum length on reaching *ACCEPT, and so ended up with the
269 wrong value of 1 rather than 0. Further investigation indicates that
270 computing a minimum subject length in the presence of *ACCEPT is difficult
271 (think back references, subroutine calls), and so I have changed the code
272 so that no minimum is registered for a pattern that contains *ACCEPT.
273
274 8. If (*THEN) was present in the first (true) branch of a conditional group,
275 it was not handled as intended. [But see 16 below.]
276
277 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
278 Sheri Pierce.
279
280 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
281 the first byte in a match must be "a".
282
283 11. Change 17 for 8.13 increased the recursion depth for patterns like
284 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
285 pattern contains any instances of (*THEN). If it does not, the old
286 optimizations are restored. It would be nice to do this on a per-group
287 basis, but at the moment that is not feasible.
288
289 12. In some environments, the output of pcretest -C is CRLF terminated. This
290 broke RunTest's code that checks for the link size. A single white space
291 character after the value is now allowed for.
292
293 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
294 For "fr", it uses the Windows-specific input and output files.
295
296 14. If (*THEN) appeared in a group that was called recursively or as a
297 subroutine, it did not work as intended. [But see next item.]
298
299 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
300 pattern fragments (but not containing any | characters). If A and B are
301 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
302 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
303 D. In other words, Perl considers parentheses that do not contain any |
304 characters to be part of a surrounding alternative, whereas PCRE was
305 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
306 differently. PCRE now behaves in the same way as Perl, except in the case
307 of subroutine/recursion calls such as (?1) which have in any case always
308 been different (but PCRE had them first :-).
309
310 16. Related to 15 above: Perl does not treat the | in a conditional group as
311 creating alternatives. Such a group is treated in the same way as an
312 ordinary group without any | characters when processing (*THEN). PCRE has
313 been changed to match Perl's behaviour.
314
315 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
316 RunGrepTest script failed.
317
318 18. Change 22 for version 13 caused atomic groups to use more stack. This is
319 inevitable for groups that contain captures, but it can lead to a lot of
320 stack use in large patterns. The old behaviour has been restored for atomic
321 groups that do not contain any capturing parentheses.
322
323 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
324 suppress the check for a minimum subject length at run time. (If it was
325 given to pcre_exec() or pcre_dfa_exec() it did work.)
326
327 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
328 fail to work when decoding hex characters in data strings in EBCDIC
329 environments.
330
331 21. It appears that in at least one Mac OS environment, the isxdigit() function
332 is implemented as a macro that evaluates to its argument more than once,
333 contravening the C 90 Standard (I haven't checked a later standard). There
334 was an instance in pcretest which caused it to go wrong when processing
335 \x{...} escapes in subject strings. The has been rewritten to avoid using
336 things like p++ in the argument of isxdigit().
337
338
339 Version 8.13 16-Aug-2011
340 ------------------------
341
342 1. The Unicode data tables have been updated to Unicode 6.0.0.
343
344 2. Two minor typos in pcre_internal.h have been fixed.
345
346 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
347 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
348 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
349
350 4. There were a number of related bugs in the code for matching backrefences
351 caselessly in UTF-8 mode when codes for the characters concerned were
352 different numbers of bytes. For example, U+023A and U+2C65 are an upper
353 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
354 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
355 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
356 2-byte code at the end of the subject (it thought there wasn't enough data
357 left).
358
359 5. Comprehensive information about what went wrong is now returned by
360 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
361 as the output vector has at least 2 elements. The offset of the start of
362 the failing character and a reason code are placed in the vector.
363
364 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
365 now returned is for the first byte of the failing character, instead of the
366 last byte inspected. This is an incompatible change, but I hope it is small
367 enough not to be a problem. It makes the returned offset consistent with
368 pcre_exec() and pcre_dfa_exec().
369
370 7. pcretest now gives a text phrase as well as the error number when
371 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
372 failure, the offset and reason code are output.
373
374 8. When \R was used with a maximizing quantifier it failed to skip backwards
375 over a \r\n pair if the subsequent match failed. Instead, it just skipped
376 back over a single character (\n). This seems wrong (because it treated the
377 two characters as a single entity when going forwards), conflicts with the
378 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
379 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
380 has been changed.
381
382 9. Some internal refactoring has changed the processing so that the handling
383 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
384 time (the PCRE_DOTALL option was changed this way some time ago: version
385 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
386 which was always a bit of a fudge. It also means that there is one less
387 argument for the match() function, which reduces its stack requirements
388 slightly. This change also fixes an incompatibility with Perl: the pattern
389 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
390
391 10. More internal refactoring has drastically reduced the number of recursive
392 calls to match() for possessively repeated groups such as (abc)++ when
393 using pcre_exec().
394
395 11. While implementing 10, a number of bugs in the handling of groups were
396 discovered and fixed:
397
398 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
399 (a|)*(?1) gave a compile-time internal error.
400 ((a|)+)+ did not notice that the outer group could match an empty string.
401 (^a|^)+ was not marked as anchored.
402 (.*a|.*)+ was not marked as matching at start or after a newline.
403
404 12. Yet more internal refactoring has removed another argument from the match()
405 function. Special calls to this function are now indicated by setting a
406 value in a variable in the "match data" data block.
407
408 13. Be more explicit in pcre_study() instead of relying on "default" for
409 opcodes that mean there is no starting character; this means that when new
410 ones are added and accidentally left out of pcre_study(), testing should
411 pick them up.
412
413 14. The -s option of pcretest has been documented for ages as being an old
414 synonym of -m (show memory usage). I have changed it to mean "force study
415 for every regex", that is, assume /S for every regex. This is similar to -i
416 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
417 using it. It makes it easier to run collections of tests with and without
418 study enabled, and thereby test pcre_study() more easily. All the standard
419 tests are now run with and without -s (but some patterns can be marked as
420 "never study" - see 20 below).
421
422 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
423 restoration of the capturing data to the outer values was not happening
424 correctly.
425
426 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
427 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
428 pattern had matched an empty string, and so incorrectly returned a no
429 match.
430
431 17. There was optimizing code for the last branch of non-capturing parentheses,
432 and also for the obeyed branch of a conditional subexpression, which used
433 tail recursion to cut down on stack usage. Unfortunately, now that there is
434 the possibility of (*THEN) occurring in these branches, tail recursion is
435 no longer possible because the return has to be checked for (*THEN). These
436 two optimizations have therefore been removed. [But see 8.20/11 above.]
437
438 18. If a pattern containing \R was studied, it was assumed that \R always
439 matched two bytes, thus causing the minimum subject length to be
440 incorrectly computed because \R can also match just one byte.
441
442 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
443 was incorrectly computed.
444
445 20. If /S is present twice on a test pattern in pcretest input, it now
446 *disables* studying, thereby overriding the use of -s on the command line
447 (see 14 above). This is necessary for one or two tests to keep the output
448 identical in both cases.
449
450 21. When (*ACCEPT) was used in an assertion that matched an empty string and
451 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
452
453 22. When an atomic group that contained a capturing parenthesis was
454 successfully matched, but the branch in which it appeared failed, the
455 capturing was not being forgotten if a higher numbered group was later
456 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
457 group 1 to "a", when in fact it should be unset. This applied to multi-
458 branched capturing and non-capturing groups, repeated or not, and also to
459 positive assertions (capturing in negative assertions does not happen
460 in PCRE) and also to nested atomic groups.
461
462 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
463 subject after a captured substring, to make it easier to tell which of a
464 number of identical substrings has been captured.
465
466 24. The way atomic groups are processed by pcre_exec() has been changed so that
467 if they are repeated, backtracking one repetition now resets captured
468 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
469 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
470 "aaa". Previously, it would have been "a". As part of this code
471 refactoring, the way recursive calls are handled has also been changed.
472
473 25. If an assertion condition captured any substrings, they were not passed
474 back unless some other capturing happened later. For example, if
475 (?(?=(a))a) was matched against "a", no capturing was returned.
476
477 26. When studying a pattern that contained subroutine calls or assertions,
478 the code for finding the minimum length of a possible match was handling
479 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
480 group 1 called group 2 while simultaneously a separate group 2 called group
481 1). A stack overflow occurred in this case. I have fixed this by limiting
482 the recursion depth to 10.
483
484 27. Updated RunTest.bat in the distribution to the version supplied by Tom
485 Fortmann. This supports explicit test numbers on the command line, and has
486 argument validation and error reporting.
487
488 28. An instance of \X with an unlimited repeat could fail if at any point the
489 first character it looked at was a mark character.
490
491 29. Some minor code refactoring concerning Unicode properties and scripts
492 should reduce the stack requirement of match() slightly.
493
494 30. Added the '=' option to pcretest to check the setting of unused capturing
495 slots at the end of the pattern, which are documented as being -1, but are
496 not included in the return count.
497
498 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
499 compiled something random. Now it gives a compile-time error (as does
500 Perl).
501
502 32. A *MARK encountered during the processing of a positive assertion is now
503 recorded and passed back (compatible with Perl).
504
505 33. If --only-matching or --colour was set on a pcregrep call whose pattern
506 had alternative anchored branches, the search for a second match in a line
507 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
508 matched the line "0102" twice. The same bug affected patterns that started
509 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
510 twice.
511
512 34. Previously, PCRE did not allow quantification of assertions. However, Perl
513 does, and because of capturing effects, quantifying parenthesized
514 assertions may at times be useful. Quantifiers are now allowed for
515 parenthesized assertions.
516
517 35. A minor code tidy in pcre_compile() when checking options for \R usage.
518
519 36. \g was being checked for fancy things in a character class, when it should
520 just be a literal "g".
521
522 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
523 appearance of a nested POSIX class supersedes an apparent external class.
524 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
525 unescaped square brackets may also appear as part of class names. For
526 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
527 more like Perl. (But see 8.20/1 above.)
528
529 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
530 was because it thought it was \N{name}, which is not supported).
531
532 39. Add minix to OS list not supporting the -S option in pcretest.
533
534 40. PCRE tries to detect cases of infinite recursion at compile time, but it
535 cannot analyze patterns in sufficient detail to catch mutual recursions
536 such as ((?1))((?2)). There is now a runtime test that gives an error if a
537 subgroup is called recursively as a subpattern for a second time at the
538 same position in the subject string. In previous releases this might have
539 been caught by the recursion limit, or it might have run out of stack.
540
541 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
542 happen only once. PCRE was, however incorrectly giving a compile time error
543 "recursive call could loop indefinitely" because it cannot analyze the
544 pattern in sufficient detail. The compile time test no longer happens when
545 PCRE is compiling a conditional subpattern, but actual runaway loops are
546 now caught at runtime (see 40 above).
547
548 42. It seems that Perl allows any characters other than a closing parenthesis
549 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
550 has been changed to be the same.
551
552 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
553 as not to get warnings when autogen.sh is called. Also changed
554 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
555
556 44. To help people who use pcregrep to scan files containing exceedingly long
557 lines, the following changes have been made:
558
559 (a) The default value of the buffer size parameter has been increased from
560 8K to 20K. (The actual buffer used is three times this size.)
561
562 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
563 PCRE is built.
564
565 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
566 to be set at run time.
567
568 (d) Numerical values in pcregrep options can be followed by K or M, for
569 example --buffer-size=50K.
570
571 (e) If a line being scanned overflows pcregrep's buffer, an error is now
572 given and the return code is set to 2.
573
574 45. Add a pointer to the latest mark to the callout data block.
575
576 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
577 partial match of an empty string instead of no match. This was specific to
578 the use of ".".
579
580 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
581 complete match instead of a partial match. This bug was dependent on both
582 the PCRE_UTF8 and PCRE_DOTALL options being set.
583
584 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
585 starting byte set, because \b was not being ignored.
586
587
588 Version 8.12 15-Jan-2011
589 ------------------------
590
591 1. Fixed some typos in the markup of the man pages, and wrote a script that
592 checks for such things as part of the documentation building process.
593
594 2. On a big-endian 64-bit system, pcregrep did not correctly process the
595 --match-limit and --recursion-limit options (added for 8.11). In
596 particular, this made one of the standard tests fail. (The integer value
597 went into the wrong half of a long int.)
598
599 3. If the --colour option was given to pcregrep with -v (invert match), it
600 did strange things, either producing crazy output, or crashing. It should,
601 of course, ignore a request for colour when reporting lines that do not
602 match.
603
604 4. Another pcregrep bug caused similar problems if --colour was specified with
605 -M (multiline) and the pattern match finished with a line ending.
606
607 5. In pcregrep, when a pattern that ended with a literal newline sequence was
608 matched in multiline mode, the following line was shown as part of the
609 match. This seems wrong, so I have changed it.
610
611 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
612 the check for further matches in the same line (so they could be coloured)
613 to overrun the end of the current line. If another match was found, it was
614 incorrectly shown (and then shown again when found in the next line).
615
616 7. If pcregrep was compiled under Windows, there was a reference to the
617 function pcregrep_exit() before it was defined. I am assuming this was
618 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
619 reported by a user. I've moved the definition above the reference.
620
621
622 Version 8.11 10-Dec-2010
623 ------------------------
624
625 1. (*THEN) was not working properly if there were untried alternatives prior
626 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
627 backtracked to try for "b" instead of moving to the next alternative branch
628 at the same level (in this case, to look for "c"). The Perl documentation
629 is clear that when (*THEN) is backtracked onto, it goes to the "next
630 alternative in the innermost enclosing group".
631
632 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
633 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
634 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
635 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
636 (*THEN).
637
638 3. If \s appeared in a character class, it removed the VT character from
639 the class, even if it had been included by some previous item, for example
640 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
641 of \s, but is part of the POSIX "space" class.)
642
643 4. A partial match never returns an empty string (because you can always
644 match an empty string at the end of the subject); however the checking for
645 an empty string was starting at the "start of match" point. This has been
646 changed to the "earliest inspected character" point, because the returned
647 data for a partial match starts at this character. This means that, for
648 example, /(?<=abc)def/ gives a partial match for the subject "abc"
649 (previously it gave "no match").
650
651 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
652 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
653 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
654 has an implication that the given string is incomplete (because a partial
655 match is preferred over a full match). For this reason, these items now
656 give a partial match in this situation. [Aside: previously, the one case
657 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
658 match rather than a full match, which was wrong by the old rules, but is
659 now correct.]
660
661 6. There was a bug in the handling of #-introduced comments, recognized when
662 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
663 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
664 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
665 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
666 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
667 places in pcre_compile().
668
669 7. Related to (6) above, when pcre_compile() was skipping #-introduced
670 comments when looking ahead for named forward references to subpatterns,
671 the only newline sequence it recognized was NL. It now handles newlines
672 according to the set newline convention.
673
674 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
675 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
676 cater for a lack of strerror(). These oversights have been fixed.
677
678 9. Added --match-limit and --recursion-limit to pcregrep.
679
680 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
681
682 11. When the -o option was used, pcregrep was setting a return code of 1, even
683 when matches were found, and --line-buffered was not being honoured.
684
685 12. Added an optional parentheses number to the -o and --only-matching options
686 of pcregrep.
687
688 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
689 can match an empty string. The code to do it in pcretest and pcredemo
690 needed fixing:
691
692 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
693 only one byte after an empty string match just before CRLF (this case
694 just got forgotten; "any" and "anycrlf" were OK).
695
696 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
697 mode when an empty string match preceded an ASCII character followed by
698 a non-ASCII character. (The code for advancing by one character rather
699 than one byte was nonsense.)
700
701 (c) The pcredemo.c sample program did not have any code at all to handle
702 the cases when CRLF is a valid newline sequence.
703
704 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
705 as a starting offset was within the subject string. There is now a new
706 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
707 negative or greater than the length of the string. In order to test this,
708 pcretest is extended to allow the setting of negative starting offsets.
709
710 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
711 starting offset points to the beginning of a UTF-8 character was
712 unnecessarily clumsy. I tidied it up.
713
714 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
715 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
716
717 17. Nobody had reported that the --include_dir option, which was added in
718 release 7.7 should have been called --include-dir (hyphen, not underscore)
719 for compatibility with GNU grep. I have changed it to --include-dir, but
720 left --include_dir as an undocumented synonym, and the same for
721 --exclude-dir, though that is not available in GNU grep, at least as of
722 release 2.5.4.
723
724 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
725 characters from a string of bytes) have been redefined so as not to use
726 loops, in order to improve performance in some environments. At the same
727 time, I abstracted some of the common code into auxiliary macros to save
728 repetition (this should not affect the compiled code).
729
730 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
731 compile-time error is now given if \c is not followed by an ASCII
732 character, that is, a byte less than 128. (In EBCDIC mode, the code is
733 different, and any byte value is allowed.)
734
735 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
736 START_OPTIMIZE option, which is now allowed at compile time - but just
737 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
738 to pcregrep and other applications that have no direct access to PCRE
739 options. The new /Y option in pcretest sets this option when calling
740 pcre_compile().
741
742 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
743 back references. Groups containing recursive back references were forced to
744 be atomic by that change, but in the case of named groups, the amount of
745 memory required was incorrectly computed, leading to "Failed: internal
746 error: code overflow". This has been fixed.
747
748 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
749 pcretest.c, to avoid build problems in some Borland environments.
750
751
752 Version 8.10 25-Jun-2010
753 ------------------------
754
755 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
756 THEN.
757
758 2. (*ACCEPT) was not working when inside an atomic group.
759
760 3. Inside a character class, \B is treated as a literal by default, but
761 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
762 causes the error). The code is unchanged, but I tidied the documentation.
763
764 4. Inside a character class, PCRE always treated \R and \X as literals,
765 whereas Perl faults them if its -w option is set. I have changed PCRE so
766 that it faults them when PCRE_EXTRA is set.
767
768 5. Added support for \N, which always matches any character other than
769 newline. (It is the same as "." when PCRE_DOTALL is not set.)
770
771 6. When compiling pcregrep with newer versions of gcc which may have
772 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
773 declared with attribute warn_unused_result" were given. Just casting the
774 result to (void) does not stop the warnings; a more elaborate fudge is
775 needed. I've used a macro to implement this.
776
777 7. Minor change to pcretest.c to avoid a compiler warning.
778
779 8. Added four artifical Unicode properties to help with an option to make
780 \s etc use properties (see next item). The new properties are: Xan
781 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
782
783 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
784 use Unicode properties. (*UCP) at the start of a pattern can be used to set
785 this option. Modified pcretest to add /W to test this facility. Added
786 REG_UCP to make it available via the POSIX interface.
787
788 10. Added --line-buffered to pcregrep.
789
790 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
791 studied, and the match started with a letter with a code point greater than
792 127 whose first byte was different to the first byte of the other case of
793 the letter, the other case of this starting letter was not recognized
794 (#976).
795
796 12. If a pattern that was studied started with a repeated Unicode property
797 test, for example, \p{Nd}+, there was the theoretical possibility of
798 setting up an incorrect bitmap of starting bytes, but fortunately it could
799 not have actually happened in practice until change 8 above was made (it
800 added property types that matched character-matching opcodes).
801
802 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
803 possible starting bytes for non-anchored patterns.
804
805 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
806 \R, and also a number of cases that involve Unicode properties, both
807 explicit and implicit when PCRE_UCP is set.
808
809 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
810 input, it could crash or give wrong results if characters with values
811 greater than 0xc0 were present in the subject string. (Detail: it assumed
812 UTF-8 input when processing these items.)
813
814 16. Added a lot of (int) casts to avoid compiler warnings in systems where
815 size_t is 64-bit (#991).
816
817 17. Added a check for running out of memory when PCRE is compiled with
818 --disable-stack-for-recursion (#990).
819
820 18. If the last data line in a file for pcretest does not have a newline on
821 the end, a newline was missing in the output.
822
823 19. The default pcre_chartables.c file recognizes only ASCII characters (values
824 less than 128) in its various bitmaps. However, there is a facility for
825 generating tables according to the current locale when PCRE is compiled. It
826 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
827 characters, are recognized by isspace() and therefore were getting set in
828 these tables, and indeed these tables seem to approximate to ISO 8859. This
829 caused a problem in UTF-8 mode when pcre_study() was used to create a list
830 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
831 which of course cannot start UTF-8 characters. I have changed the code so
832 that only real ASCII characters (less than 128) and the correct starting
833 bytes for UTF-8 encodings are set for characters greater than 127 when in
834 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
835 altogether.)
836
837 20. Added the /T option to pcretest so as to be able to run tests with non-
838 standard character tables, thus making it possible to include the tests
839 used for 19 above in the standard set of tests.
840
841 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
842 reference to a subpattern the other side of a comment that contains an
843 opening parenthesis caused either an internal compiling error, or a
844 reference to the wrong subpattern.
845
846
847 Version 8.02 19-Mar-2010
848 ------------------------
849
850 1. The Unicode data tables have been updated to Unicode 5.2.0.
851
852 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
853 configured.
854
855 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
856 original author of that file, following a query about its status.
857
858 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
859 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
860
861 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
862 quantifier applied to a forward-referencing subroutine call, could compile
863 incorrect code or give the error "internal error: previously-checked
864 referenced subpattern not found".
865
866 6. Both MS Visual Studio and Symbian OS have problems with initializing
867 variables to point to external functions. For these systems, therefore,
868 pcre_malloc etc. are now initialized to local functions that call the
869 relevant global functions.
870
871 7. There were two entries missing in the vectors called coptable and poptable
872 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
873 I've fixed the data, and added a kludgy way of testing at compile time that
874 the lengths are correct (equal to the number of opcodes).
875
876 8. Following on from 7, I added a similar kludge to check the length of the
877 eint vector in pcreposix.c.
878
879 9. Error texts for pcre_compile() are held as one long string to avoid too
880 much relocation at load time. To find a text, the string is searched,
881 counting zeros. There was no check for running off the end of the string,
882 which could happen if a new error number was added without updating the
883 string.
884
885 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
886
887 11. \K was not working if it appeared in an atomic group or in a group that
888 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
889 \K is "not well defined" if used in an assertion. PCRE now accepts it if
890 the assertion is positive, but not if it is negative.
891
892 12. Change 11 fortuitously reduced the size of the stack frame used in the
893 "match()" function of pcre_exec.c by one pointer. Forthcoming
894 implementation of support for (*MARK) will need an extra pointer on the
895 stack; I have reserved it now, so that the stack frame size does not
896 decrease.
897
898 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
899 item in branch that calls a recursion is a subroutine call - as in the
900 second branch in the above example - was incorrectly given the compile-
901 time error "recursive call could loop indefinitely" because pcre_compile()
902 was not correctly checking the subroutine for matching a non-empty string.
903
904 14. The checks for overrunning compiling workspace could trigger after an
905 overrun had occurred. This is a "should never occur" error, but it can be
906 triggered by pathological patterns such as hundreds of nested parentheses.
907 The checks now trigger 100 bytes before the end of the workspace.
908
909 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
910
911
912 Version 8.01 19-Jan-2010
913 ------------------------
914
915 1. If a pattern contained a conditional subpattern with only one branch (in
916 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
917 computed the wrong minimum data length (which is of course zero for such
918 subpatterns). This could cause incorrect "no match" results.
919
920 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
921 the pattern is reset in the first branch, pcre_compile() failed with
922 "internal error: code overflow at offset...". This happened only when
923 the reset was to the original external option setting. (An optimization
924 abstracts leading options settings into an external setting, which was the
925 cause of this.)
926
927 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
928 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
929 assertion pattern did not match (meaning that the assertion was true), it
930 was incorrectly treated as false if the SKIP had been reached during the
931 matching. This also applied to assertions used as conditions.
932
933 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
934 assertion subpattern, including such a pattern used as a condition,
935 unpredictable results occurred, instead of the error return
936 PCRE_ERROR_DFA_UITEM.
937
938 5. The C++ GlobalReplace function was not working like Perl for the special
939 situation when an empty string is matched. It now does the fancy magic
940 stuff that is necessary.
941
942 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
943 removed. (These were left over from very, very early versions of PCRE.)
944
945 7. Some cosmetic changes to the code to make life easier when compiling it
946 as part of something else:
947
948 (a) Change DEBUG to PCRE_DEBUG.
949
950 (b) In pcre_compile(), rename the member of the "branch_chain" structure
951 called "current" as "current_branch", to prevent a collision with the
952 Linux macro when compiled as a kernel module.
953
954 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
955 prevent a collision with the Linux macro when compiled as a kernel
956 module.
957
958 8. In pcre_compile() there are some checks for integer overflows that used to
959 cast potentially large values to (double). This has been changed to that
960 when building, a check for int64_t is made, and if it is found, it is used
961 instead, thus avoiding the use of floating point arithmetic. (There is no
962 other use of FP in PCRE.) If int64_t is not found, the fallback is to
963 double.
964
965 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
966 2005 (difference between two addresses compared to an unsigned value).
967
968 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
969 custom one, because of the following reported problem in Windows:
970
971 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
972 under Win32.
973 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
974 therefore missing the function definition.
975 - The compiler thus generates a "C" signature for the test function.
976 - The linker fails to find the "C" function.
977 - PCRE fails to configure if asked to do so against libbz2.
978
979 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
980 messages were output:
981
982 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
983 rerunning libtoolize, to keep the correct libtool macros in-tree.
984 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
985
986 I have done both of these things.
987
988 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
989 most of the time, it *can* run out if it is given a pattern that contains a
990 runaway infinite recursion. I updated the discussion in the pcrestack man
991 page.
992
993 13. Now that we have gone to the x.xx style of version numbers, the minor
994 version may start with zero. Using 08 or 09 is a bad idea because users
995 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
996 interpreted as invalid octal numbers. I've updated the previous comment in
997 configure.ac, and also added a check that gives an error if 08 or 09 are
998 used.
999
1000 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1001 causing partial matching to fail when the end of the subject matched \W
1002 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1003
1004 15. There were some discrepancies between the declarations in pcre_internal.h
1005 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1006 their definitions. The declarations used "const uschar *" and the
1007 definitions used USPTR. Even though USPTR is normally defined as "const
1008 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1009 reported that: "This difference in casting confuses some C++ compilers, for
1010 example, SunCC recognizes above declarations as different functions and
1011 generates broken code for hbpcre." I have changed the declarations to use
1012 USPTR.
1013
1014 16. GNU libtool is named differently on some systems. The autogen.sh script now
1015 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1016 (FreeBSD).
1017
1018 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1019 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1020 comment: "Figure out how to create a longlong from a string: strtoll and
1021 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1022 instance, but it only takes 2 args instead of 3!"
1023
1024 18. A subtle bug concerned with back references has been fixed by a change of
1025 specification, with a corresponding code fix. A pattern such as
1026 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1027 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1028 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1029 same bug. Such groups have to be quantified to be useful, or contained
1030 inside another quantified group. (If there's no repetition, the reference
1031 can never match.) The problem arises because, having left the group and
1032 moved on to the rest of the pattern, a later failure that backtracks into
1033 the group uses the captured value from the final iteration of the group
1034 rather than the correct earlier one. I have fixed this in PCRE by forcing
1035 any group that contains a reference to itself to be an atomic group; that
1036 is, there cannot be any backtracking into it once it has completed. This is
1037 similar to recursive and subroutine calls.
1038
1039
1040 Version 8.00 19-Oct-09
1041 ----------------------
1042
1043 1. The table for translating pcre_compile() error codes into POSIX error codes
1044 was out-of-date, and there was no check on the pcre_compile() error code
1045 being within the table. This could lead to an OK return being given in
1046 error.
1047
1048 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1049 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1050 in a Windows environment.
1051
1052 3. The pcregrep --count option prints the count for each file even when it is
1053 zero, as does GNU grep. However, pcregrep was also printing all files when
1054 --files-with-matches was added. Now, when both options are given, it prints
1055 counts only for those files that have at least one match. (GNU grep just
1056 prints the file name in this circumstance, but including the count seems
1057 more useful - otherwise, why use --count?) Also ensured that the
1058 combination -clh just lists non-zero counts, with no names.
1059
1060 4. The long form of the pcregrep -F option was incorrectly implemented as
1061 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1062 but it seems right to fix it, and I didn't think it was worth preserving
1063 the old behaviour.
1064
1065 5. The command line items --regex=pattern and --regexp=pattern were not
1066 recognized by pcregrep, which required --regex pattern or --regexp pattern
1067 (with a space rather than an '='). The man page documented the '=' forms,
1068 which are compatible with GNU grep; these now work.
1069
1070 6. No libpcreposix.pc file was created for pkg-config; there was just
1071 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1072
1073 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1074 when UCP support is not needed, by modifying the Python script that
1075 generates it from Unicode data files. This should not matter if the module
1076 is correctly used as a library, but I received one complaint about 50K of
1077 unwanted data. My guess is that the person linked everything into his
1078 program rather than using a library. Anyway, it does no harm.
1079
1080 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1081 was a minimum greater than 1 for a wide character in a possessive
1082 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1083 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1084 character. Chaos in the form of incorrect output or a compiling loop could
1085 result.
1086
1087 9. The restrictions on what a pattern can contain when partial matching is
1088 requested for pcre_exec() have been removed. All patterns can now be
1089 partially matched by this function. In addition, if there are at least two
1090 slots in the offset vector, the offset of the earliest inspected character
1091 for the match and the offset of the end of the subject are set in them when
1092 PCRE_ERROR_PARTIAL is returned.
1093
1094 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1095 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1096 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1097 and may be more useful for multi-segment matching.
1098
1099 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1100 used to be given if ever the end of the subject was reached; now it is
1101 given only if matching could not proceed because another character was
1102 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1103 string "Z", which now yields "no match" instead of "partial match". In the
1104 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1105 final character ended with (*FAIL).
1106
1107 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1108 if the pattern had a "must contain" character that was already found in the
1109 earlier partial match, unless partial matching was again requested. For
1110 example, with the pattern /dog.(body)?/, the "must contain" character is
1111 "g". If the first part-match was for the string "dog", restarting with
1112 "sbody" failed. This bug has been fixed.
1113
1114 13. The string returned by pcre_dfa_exec() after a partial match has been
1115 changed so that it starts at the first inspected character rather than the
1116 first character of the match. This makes a difference only if the pattern
1117 starts with a lookbehind assertion or \b or \B (\K is not supported by
1118 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1119 matching functions compatible, and I think it's the right thing to do.
1120
1121 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1122 so that the demonstration program is easily available in environments where
1123 PCRE has not been installed from source.
1124
1125 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1126 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1127 library.
1128
1129 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1130 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1131 is not the first non-POSIX option to be added. Clearly some people find
1132 these options useful.
1133
1134 17. If a caller to the POSIX matching function regexec() passes a non-zero
1135 value for nmatch with a NULL value for pmatch, the value of
1136 nmatch is forced to zero.
1137
1138 18. RunGrepTest did not have a test for the availability of the -u option of
1139 the diff command, as RunTest does. It now checks in the same way as
1140 RunTest, and also checks for the -b option.
1141
1142 19. If an odd number of negated classes containing just a single character
1143 interposed, within parentheses, between a forward reference to a named
1144 subpattern and the definition of the subpattern, compilation crashed with
1145 an internal error, complaining that it could not find the referenced
1146 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1147 [The bug was that it was starting one character too far in when skipping
1148 over the character class, thus treating the ] as data rather than
1149 terminating the class. This meant it could skip too much.]
1150
1151 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1152 /g option in pcretest when the pattern contains \K, which makes it possible
1153 to have an empty string match not at the start, even when the pattern is
1154 anchored. Updated pcretest and pcredemo to use this option.
1155
1156 21. If the maximum number of capturing subpatterns in a recursion was greater
1157 than the maximum at the outer level, the higher number was returned, but
1158 with unset values at the outer level. The correct (outer level) value is
1159 now given.
1160
1161 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1162 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1163 make it do so. The string so far is captured, making this feature
1164 compatible with Perl.
1165
1166 23. The tests have been re-organized, adding tests 11 and 12, to make it
1167 possible to check the Perl 5.10 features against Perl 5.10.
1168
1169 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1170 pattern matches a fixed length string. PCRE did not allow this; now it
1171 does. Neither allows recursion.
1172
1173 25. I finally figured out how to implement a request to provide the minimum
1174 length of subject string that was needed in order to match a given pattern.
1175 (It was back references and recursion that I had previously got hung up
1176 on.) This code has now been added to pcre_study(); it finds a lower bound
1177 to the length of subject needed. It is not necessarily the greatest lower
1178 bound, but using it to avoid searching strings that are too short does give
1179 some useful speed-ups. The value is available to calling programs via
1180 pcre_fullinfo().
1181
1182 26. While implementing 25, I discovered to my embarrassment that pcretest had
1183 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1184 study optimizations had never been tested with that matching function.
1185 Oops. What is worse, even when it was passed study data, there was a bug in
1186 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1187 were also very few tests of studied patterns with pcre_dfa_exec().
1188
1189 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1190 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1191 on the other side of the coin, they are no longer allowed to have different
1192 names, because these cannot be distinguished in PCRE, and this has caused
1193 confusion. (This is a difference from Perl.)
1194
1195 28. When duplicate subpattern names are present (necessarily with different
1196 numbers, as required by 27 above), and a test is made by name in a
1197 conditional pattern, either for a subpattern having been matched, or for
1198 recursion in such a pattern, all the associated numbered subpatterns are
1199 tested, and the overall condition is true if the condition is true for any
1200 one of them. This is the way Perl works, and is also more like the way
1201 testing by number works.
1202
1203
1204 Version 7.9 11-Apr-09
1205 ---------------------
1206
1207 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1208 (pcretest), all targets were linked against these libraries. This included
1209 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1210 libraries. This caused unwanted dependencies to be created. This problem
1211 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1212 pcretest is linked with readline.
1213
1214 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1215 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1216 moved outside it again, because FALSE and TRUE are already defined in AIX,
1217 but BOOL is not.
1218
1219 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1220 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1221
1222 4. The pcregrep documentation said spaces were inserted as well as colons (or
1223 hyphens) following file names and line numbers when outputting matching
1224 lines. This is not true; no spaces are inserted. I have also clarified the
1225 wording for the --colour (or --color) option.
1226
1227 5. In pcregrep, when --colour was used with -o, the list of matching strings
1228 was not coloured; this is different to GNU grep, so I have changed it to be
1229 the same.
1230
1231 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1232 each matching line was coloured. Now it goes on to look for further matches
1233 of any of the test patterns, which is the same behaviour as GNU grep.
1234
1235 7. A pattern that could match an empty string could cause pcregrep to loop; it
1236 doesn't make sense to accept an empty string match in pcregrep, so I have
1237 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1238 seems to be how GNU grep behaves.
1239
1240 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1241 start or after a newline", because the conditional assertion was not being
1242 correctly handled. The rule now is that both the assertion and what follows
1243 in the first alternative must satisfy the test.
1244
1245 9. If auto-callout was enabled in a pattern with a conditional group whose
1246 condition was an assertion, PCRE could crash during matching, both with
1247 pcre_exec() and pcre_dfa_exec().
1248
1249 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1250 used for matching.
1251
1252 11. Unicode property support in character classes was not working for
1253 characters (bytes) greater than 127 when not in UTF-8 mode.
1254
1255 12. Added the -M command line option to pcretest.
1256
1257 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1258
1259 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1260
1261 16. Added comments and documentation about mis-use of no_arg in the C++
1262 wrapper.
1263
1264 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1265 from Martin Jerabek that uses macro names for all relevant character and
1266 string constants.
1267
1268 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1269 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1270 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1271 these, but not everybody uses configure.
1272
1273 19. A conditional group that had only one branch was not being correctly
1274 recognized as an item that could match an empty string. This meant that an
1275 enclosing group might also not be so recognized, causing infinite looping
1276 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1277 with the subject "ab", where knowledge that the repeated group can match
1278 nothing is needed in order to break the loop.
1279
1280 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1281 exec(), but without supplying a callout function, matching went wrong.
1282
1283 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1284 leak if the size of the offset vector was greater than 30. When the vector
1285 is smaller, the saved offsets during recursion go onto a local stack
1286 vector, but for larger vectors malloc() is used. It was failing to free
1287 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1288 error, in fact).
1289
1290 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1291 heapframe that is used only when UTF-8 support is enabled. This caused no
1292 problem, but was untidy.
1293
1294 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1295 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1296 included within another project.
1297
1298 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1299 slightly modified by me:
1300
1301 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1302 not building pcregrep.
1303
1304 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1305 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1306
1307 25. Forward references, both numeric and by name, in patterns that made use of
1308 duplicate group numbers, could behave incorrectly or give incorrect errors,
1309 because when scanning forward to find the reference group, PCRE was not
1310 taking into account the duplicate group numbers. A pattern such as
1311 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1312
1313 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1314 the feature of a custom pointer more persuasive (as requested by a user).
1315
1316 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1317 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1318 pre-defined.
1319
1320 28. Added support for (*UTF8) at the start of a pattern.
1321
1322 29. Arrange for flags added by the "release type" setting in CMake to be shown
1323 in the configuration summary.
1324
1325
1326 Version 7.8 05-Sep-08
1327 ---------------------
1328
1329 1. Replaced UCP searching code with optimized version as implemented for Ad
1330 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1331 stage table and inline lookup instead of a function, giving speed ups of 2
1332 to 5 times on some simple patterns that I tested. Permission was given to
1333 distribute the MultiStage2.py script that generates the tables (it's not in
1334 the tarball, but is in the Subversion repository).
1335
1336 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1337 scripts.
1338
1339 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1340 a group with a zero qualifier. The result of the study could be incorrect,
1341 or the function might crash, depending on the pattern.
1342
1343 4. Caseless matching was not working for non-ASCII characters in back
1344 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1345 It now works when Unicode Property Support is available.
1346
1347 5. In pcretest, an escape such as \x{de} in the data was always generating
1348 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1349 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1350 truncation.
1351
1352 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1353
1354 7. Added two (int) casts to pcregrep when printing the difference of two
1355 pointers, in case they are 64-bit values.
1356
1357 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1358 test 2 if it fails.
1359
1360 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1361 and a #define of that name to empty if it is not externally set. This is to
1362 allow users of MSVC to set it if necessary.
1363
1364 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1365 the convenience functions in the pcre_get.c source file.
1366
1367 11. An option change at the start of a pattern that had top-level alternatives
1368 could cause overwriting and/or a crash. This command provoked a crash in
1369 some environments:
1370
1371 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1372
1373 This potential security problem was recorded as CVE-2008-2371.
1374
1375 12. For a pattern where the match had to start at the beginning or immediately
1376 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1377 pcre_dfa_exec() could read past the end of the passed subject if there was
1378 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1379 pcretest so that it places the subject at the end of its malloc-ed buffer.
1380
1381 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1382 exec() might read past the end of the data buffer in UTF-8 mode.
1383
1384 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1385 the data contained the byte 0x85 as part of a UTF-8 character within its
1386 first line. This applied both to normal and DFA matching.
1387
1388 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1389 /^[^d]*?$/8 failed to match "abc".
1390
1391 16. Added a missing copyright notice to pcrecpp_internal.h.
1392
1393 17. Make it more clear in the documentation that values returned from
1394 pcre_exec() in ovector are byte offsets, not character counts.
1395
1396 18. Tidied a few places to stop certain compilers from issuing warnings.
1397
1398 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1399 supplied by Stefan Weber. I made a further small update for 7.8 because
1400 there is a change of source arrangements: the pcre_searchfuncs.c module is
1401 replaced by pcre_ucd.c.
1402
1403
1404 Version 7.7 07-May-08
1405 ---------------------
1406
1407 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1408 a string to a long long, pretend we don't even have a long long." This is
1409 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1410
1411 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1412 pre-7.6 versions, which defined a global no_arg variable instead of putting
1413 it in the RE class. (See also #8 below.)
1414
1415 3. Remove a line of dead code, identified by coverity and reported by Nuno
1416 Lopes.
1417
1418 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1419
1420 (1) The include/exclude patterns were being applied to the whole pathnames
1421 of files, instead of just to the final components.
1422
1423 (2) If there was more than one level of directory, the subdirectories were
1424 skipped unless they satisfied the include/exclude conditions. This is
1425 inconsistent with GNU grep (and could even be seen as contrary to the
1426 pcregrep specification - which I improved to make it absolutely clear).
1427 The action now is always to scan all levels of directory, and just
1428 apply the include/exclude patterns to regular files.
1429
1430 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1431 --exclude_dir in the tests to avoid scanning .svn directories.
1432
1433 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1434 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1435 doesn't support NULs in patterns.
1436
1437 7. Added some missing "const"s to declarations of static tables in
1438 pcre_compile.c and pcre_dfa_exec.c.
1439
1440 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1441 caused by fix #2 above. (Subsequently also a second patch to fix the
1442 first patch. And a third patch - this was a messy problem.)
1443
1444 9. Applied Craig's patch to remove the use of push_back().
1445
1446 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1447 matching function regexec().
1448
1449 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1450 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1451 references. PCRE supports relative numbers with this syntax (I don't think
1452 Oniguruma does).
1453
1454 12. Previously, a group with a zero repeat such as (...){0} was completely
1455 omitted from the compiled regex. However, this means that if the group
1456 was called as a subroutine from elsewhere in the pattern, things went wrong
1457 (an internal error was given). Such groups are now left in the compiled
1458 pattern, with a new opcode that causes them to be skipped at execution
1459 time.
1460
1461 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1462 to the way PCRE behaves:
1463
1464 (a) A lone ] character is dis-allowed (Perl treats it as data).
1465
1466 (b) A back reference to an unmatched subpattern matches an empty string
1467 (Perl fails the current match path).
1468
1469 (c) A data ] in a character class must be notated as \] because if the
1470 first data character in a class is ], it defines an empty class. (In
1471 Perl it is not possible to have an empty class.) The empty class []
1472 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1473 The negative empty class [^] matches any one character, independently
1474 of the DOTALL setting.
1475
1476 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1477 non-existent subpattern following a character class starting with ']' and
1478 containing () gave an internal compiling error instead of "reference to
1479 non-existent subpattern". Fortunately, when the pattern did exist, the
1480 compiled code was correct. (When scanning forwards to check for the
1481 existencd of the subpattern, it was treating the data ']' as terminating
1482 the class, so got the count wrong. When actually compiling, the reference
1483 was subsequently set up correctly.)
1484
1485 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1486 it was being rejected as not supported by pcre_dfa_exec(), even though
1487 other assertions are supported. I have made pcre_dfa_exec() support
1488 (*FAIL).
1489
1490 16. The implementation of 13c above involved the invention of a new opcode,
1491 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1492 cannot be changed at match time, I realized I could make a small
1493 improvement to matching performance by compiling OP_ALLANY instead of
1494 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1495 on the OP_ANY path.
1496
1497 17. Compiling pcretest on Windows with readline support failed without the
1498 following two fixes: (1) Make the unistd.h include conditional on
1499 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1500
1501 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1502 ncurses library to be included for pcretest when ReadLine support is
1503 requested, but also to allow for it to be overridden. This patch came from
1504 Daniel Bergström.
1505
1506 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1507 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1508 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1509 spotting this.
1510
1511
1512 Version 7.6 28-Jan-08
1513 ---------------------
1514
1515 1. A character class containing a very large number of characters with
1516 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1517 overflow.
1518
1519 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1520 HAVE_LONG_LONG is not defined.
1521
1522 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1523 bring them up to date and include new features. This patch includes:
1524
1525 - Fixed PH's badly added libz and libbz2 support.
1526 - Fixed a problem with static linking.
1527 - Added pcredemo. [But later removed - see 7 below.]
1528 - Fixed dftables problem and added an option.
1529 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1530 HAVE_LONG_LONG.
1531 - Added readline support for pcretest.
1532 - Added an listing of the option settings after cmake has run.
1533
1534 4. A user submitted a patch to Makefile that makes it easy to create
1535 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1536 Makefile.am that cause it to include this special target, without
1537 affecting anything else. Note that the same mingw target plus all
1538 the other distribution libraries and programs are now supported
1539 when configuring with CMake (see 6 below) instead of with
1540 Configure/Make.
1541
1542 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1543 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1544 exported in the Windows port". It has not yet been confirmed that the patch
1545 solves the problem, but it does no harm.
1546
1547 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1548 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1549 with CMake, and also correct the comment about stack recursion.
1550
1551 7. Remove the automatic building of pcredemo from the ./configure system and
1552 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1553 of a program that users should build themselves after PCRE is installed, so
1554 building it automatically is not really right. What is more, it gave
1555 trouble in some build environments.
1556
1557 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1558
1559
1560 Version 7.5 10-Jan-08
1561 ---------------------
1562
1563 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1564 values in parens when parsing an RE using the C++ wrapper."
1565
1566 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1567 Characters greater than 255 were excluded from the class instead of being
1568 included.
1569
1570 3. The same bug as (2) above applied to negated POSIX classes such as
1571 [:^space:].
1572
1573 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1574 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1575 I have changed it.
1576
1577 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1578 first named subpattern) and a construct such as (?&a) would reference the
1579 first named subpattern whose name started with "a" (in other words, the
1580 length check was missing). Both these problems are fixed. "Subpattern name
1581 expected" is now given for (?&) (a zero-length name), and this patch also
1582 makes it give the same error for \k'' (previously it complained that that
1583 was a reference to a non-existent subpattern).
1584
1585 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1586 this is right because (?- can be followed by option settings as well as by
1587 digits. I have, however, made the messages clearer.
1588
1589 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1590 than the number used in the conditional) now cause a compile-time error.
1591 This is actually not compatible with Perl, which accepts such patterns, but
1592 treats the conditional as always being FALSE (as PCRE used to), but it
1593 seems to me that giving a diagnostic is better.
1594
1595 8. Change "alphameric" to the more common word "alphanumeric" in comments
1596 and messages.
1597
1598 9. Fix two occurrences of "backslash" in comments that should have been
1599 "backspace".
1600
1601 10. Remove two redundant lines of code that can never be obeyed (their function
1602 was moved elsewhere).
1603
1604 11. The program that makes PCRE's Unicode character property table had a bug
1605 which caused it to generate incorrect table entries for sequences of
1606 characters that have the same character type, but are in different scripts.
1607 It amalgamated them into a single range, with the script of the first of
1608 them. In other words, some characters were in the wrong script. There were
1609 thirteen such cases, affecting characters in the following ranges:
1610
1611 U+002b0 - U+002c1
1612 U+0060c - U+0060d
1613 U+0061e - U+00612
1614 U+0064b - U+0065e
1615 U+0074d - U+0076d
1616 U+01800 - U+01805
1617 U+01d00 - U+01d77
1618 U+01d9b - U+01dbf
1619 U+0200b - U+0200f
1620 U+030fc - U+030fe
1621 U+03260 - U+0327f
1622 U+0fb46 - U+0fbb1
1623 U+10450 - U+1049d
1624
1625 12. The -o option (show only the matching part of a line) for pcregrep was not
1626 compatible with GNU grep in that, if there was more than one match in a
1627 line, it showed only the first of them. It now behaves in the same way as
1628 GNU grep.
1629
1630 13. If the -o and -v options were combined for pcregrep, it printed a blank
1631 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1632 does the same. The return code can be used to tell if there were any
1633 non-matching lines.
1634
1635 14. Added --file-offsets and --line-offsets to pcregrep.
1636
1637 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1638 infinitely looping recursion. The bug was that positive lookaheads were not
1639 being skipped when checking for a possible empty match (negative lookaheads
1640 and both kinds of lookbehind were skipped).
1641
1642 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1643 inclusion of <windows.h> to before rather than after the definition of
1644 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1645
1646 17. Specifying a possessive quantifier with a specific limit for a Unicode
1647 character property caused pcre_compile() to compile bad code, which led at
1648 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1649 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1650 caused the error; without that there was no problem.
1651
1652 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1653
1654 19. Added --enable-pcretest-libreadline.
1655
1656 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1657 RE::GlobalReplace(). As a result, the number of replacements returned was
1658 double what it should be. I removed one of the increments, but Craig sent a
1659 later patch that removed the other one (the right fix) and added unit tests
1660 that check the return values (which was not done before).
1661
1662 21. Several CMake things:
1663
1664 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1665 the names libpcre and libpcreposix, not just pcre and pcreposix.
1666
1667 (2) The above change means that pcretest and pcregrep are now correctly
1668 linked with the newly-built libraries, not previously installed ones.
1669
1670 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1671
1672 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1673 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1674 UTF-8 newline character). The key issue is that the pattern starts .*;
1675 this means that the match must be either at the beginning, or after a
1676 newline. The bug was in the code for advancing after a failed match and
1677 checking that the new position followed a newline. It was not taking
1678 account of UTF-8 characters correctly.
1679
1680 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1681 character classes. PCRE was not treating the sequence [:...:] as a
1682 character class unless the ... were all letters. Perl, however, seems to
1683 allow any characters between [: and :], though of course it rejects as
1684 unknown any "names" that contain non-letters, because all the known class
1685 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1686 for example, whereas PCRE did not - it did not recognize a POSIX character
1687 class. This seemed a bit dangerous, so the code has been changed to be
1688 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1689 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1690 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1691 Perl does, and where it didn't before.
1692
1693 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1694 Windows environments %n is disabled by default.
1695
1696
1697 Version 7.4 21-Sep-07
1698 ---------------------
1699
1700 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1701 means that a class such as [\s] counted as "explicit reference to CR or
1702 LF". That isn't really right - the whole point of the change was to try to
1703 help when there was an actual mention of one of the two characters. So now
1704 the change happens only if \r or \n (or a literal CR or LF) character is
1705 encountered.
1706
1707 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1708 of both had grown to the point where there were only 3 bits left.
1709 Fortunately, there was spare space in the data structure, and so I have
1710 moved the internal flags into a new 16-bit field to free up more option
1711 bits.
1712
1713 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1714 but did not set the internal JCHANGED flag - either of these is enough to
1715 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1716 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1717 start sets both bits.
1718
1719 4. Added options (at build time, compile time, exec time) to change \R from
1720 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1721
1722 5. doc/pcresyntax.html was missing from the distribution.
1723
1724 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1725 compatibility, even though it is no longer used.
1726
1727 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1728 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1729 windows.h file is present (where different names are used). [This was
1730 reversed later after testing - see 16 below.]
1731
1732 8. Changed all #include <config.h> to #include "config.h". There were also
1733 some further <pcre.h> cases that I changed to "pcre.h".
1734
1735 9. When pcregrep was used with the --colour option, it missed the line ending
1736 sequence off the lines that it output.
1737
1738 10. It was pointed out to me that arrays of string pointers cause lots of
1739 relocations when a shared library is dynamically loaded. A technique of
1740 using a single long string with a table of offsets can drastically reduce
1741 these. I have refactored PCRE in four places to do this. The result is
1742 dramatic:
1743
1744 Originally: 290
1745 After changing UCP table: 187
1746 After changing error message table: 43
1747 After changing table of "verbs" 36
1748 After changing table of Posix names 22
1749
1750 Thanks to the folks working on Gregex for glib for this insight.
1751
1752 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1753 unicode-properties was also set.
1754
1755 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1756
1757 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1758 checked only for CRLF.
1759
1760 14. Added casts to pcretest.c to avoid compiler warnings.
1761
1762 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1763
1764 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1765 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1766 entirely. This removes changes made in 7 above.
1767
1768 17. The CMake files have been updated, and there is now more information about
1769 building with CMake in the NON-UNIX-USE document.
1770
1771
1772 Version 7.3 28-Aug-07
1773 ---------------------
1774
1775 1. In the rejigging of the build system that eventually resulted in 7.1, the
1776 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1777 brackets there is not right, since it causes compilers to look for an
1778 installed pcre.h, not the version that is in the source that is being
1779 compiled (which of course may be different). I have changed it back to:
1780
1781 #include "pcre.h"
1782
1783 I have a vague recollection that the change was concerned with compiling in
1784 different directories, but in the new build system, that is taken care of
1785 by the VPATH setting the Makefile.
1786
1787 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1788 when the subject happened to end in the byte 0x85 (e.g. if the last
1789 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1790 characters but of course it shouldn't be taken as a newline when it is part
1791 of another character. The bug was that, for an unlimited repeat of . in
1792 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1793 characters when looking for a newline.
1794
1795 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1796
1797 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1798 in debug output.
1799
1800 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1801 long printing in the pcrecpp unittest when running under MinGW.
1802
1803 6. ESC_K was left out of the EBCDIC table.
1804
1805 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1806 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1807 limit also applies to "virtual nesting" when a pattern is recursive, and in
1808 this case 1000 isn't so big. I have been able to remove this limit at the
1809 expense of backing off one optimization in certain circumstances. Normally,
1810 when pcre_exec() would call its internal match() function recursively and
1811 immediately return the result unconditionally, it uses a "tail recursion"
1812 feature to save stack. However, when a subpattern that can match an empty
1813 string has an unlimited repetition quantifier, it no longer makes this
1814 optimization. That gives it a stack frame in which to save the data for
1815 checking that an empty string has been matched. Previously this was taken
1816 from the 1000-entry workspace that had been reserved. So now there is no
1817 explicit limit, but more stack is used.
1818
1819 8. Applied Daniel's patches to solve problems with the import/export magic
1820 syntax that is required for Windows, and which was going wrong for the
1821 pcreposix and pcrecpp parts of the library. These were overlooked when this
1822 problem was solved for the main library.
1823
1824 9. There were some crude static tests to avoid integer overflow when computing
1825 the size of patterns that contain repeated groups with explicit upper
1826 limits. As the maximum quantifier is 65535, the maximum group length was
1827 set at 30,000 so that the product of these two numbers did not overflow a
1828 32-bit integer. However, it turns out that people want to use groups that
1829 are longer than 30,000 bytes (though not repeat them that many times).
1830 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1831 made it possible to implement the integer overflow checks in a much more
1832 dynamic way, which I have now done. The artificial limitation on group
1833 length has been removed - we now have only the limit on the total length of
1834 the compiled pattern, which depends on the LINK_SIZE setting.
1835
1836 10. Fixed a bug in the documentation for get/copy named substring when
1837 duplicate names are permitted. If none of the named substrings are set, the
1838 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1839 empty string.
1840
1841 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1842 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1843 because the ] is interpreted as the first data character and the
1844 terminating ] is not found. PCRE has been made compatible with Perl in this
1845 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1846 cause memory overwriting.
1847
1848 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1849 string has been matched (to stop an infinite loop). It was not recognizing
1850 a conditional subpattern that could match an empty string if that
1851 subpattern was within another subpattern. For example, it looped when
1852 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1853 condition was not nested. This bug has been fixed.
1854
1855 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1856 past the start of the subject in the presence of bytes with the top bit
1857 set, for example "\x8aBCD".
1858
1859 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1860 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1861
1862 14. Optimized (?!) to (*FAIL).
1863
1864 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1865 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1866 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1867 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1868 does: it's just the validity check that is more restrictive.
1869
1870 16. Inserted checks for integer overflows during escape sequence (backslash)
1871 processing, and also fixed erroneous offset values for syntax errors during
1872 backslash processing.
1873
1874 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1875 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1876
1877 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1878 caused an overrun.
1879
1880 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1881 something other than just ASCII characters) inside a group that had an
1882 unlimited repeat caused a loop at compile time (while checking to see
1883 whether the group could match an empty string).
1884
1885 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1886 [\P{Any}] did so. (Error in the code for printing property names.)
1887
1888 21. An orphan \E inside a character class could cause a crash.
1889
1890 22. A repeated capturing bracket such as (A)? could cause a wild memory
1891 reference during compilation.
1892
1893 23. There are several functions in pcre_compile() that scan along a compiled
1894 expression for various reasons (e.g. to see if it's fixed length for look
1895 behind). There were bugs in these functions when a repeated \p or \P was
1896 present in the pattern. These operators have additional parameters compared
1897 with \d, etc, and these were not being taken into account when moving along
1898 the compiled data. Specifically:
1899
1900 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1901 length.
1902
1903 (b) An item such as \pL+ within a repeated group could cause crashes or
1904 loops.
1905
1906 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1907 "reference to non-existent subpattern" error.
1908
1909 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1910
1911 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1912 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1913
1914 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1915
1916 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1917 character were causing crashes (broken optimization).
1918
1919 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1920 \p or \P) caused a compile-time loop.
1921
1922 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1923 break. For example, the unstudied pattern [\r\n]A does not match the string
1924 "\r\nA" because change 7.0/46 below moves the current point on by two
1925 characters after failing to match at the start. However, the pattern \nA
1926 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1927 the same is true. There doesn't seem any very clean way out of this, but
1928 what I have chosen to do makes the common cases work: PCRE now takes note
1929 of whether there can be an explicit match for \r or \n anywhere in the
1930 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1931 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1932 pattern has explicit CR or LF references.
1933
1934 29. Added (*CR) etc for changing newline setting at start of pattern.
1935
1936
1937 Version 7.2 19-Jun-07
1938 ---------------------
1939
1940 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1941 which is apparently normally available under Windows.
1942
1943 2. Re-jig the pcregrep tests with different newline settings in an attempt
1944 to make them independent of the local environment's newline setting.
1945
1946 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1947
1948 4. Some of the "internals" tests were previously cut out when the link size
1949 was not 2, because the output contained actual offsets. The recent new
1950 "Z" feature of pcretest means that these can be cut out, making the tests
1951 usable with all link sizes.
1952
1953 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1954 stack recursion. This gives a massive performance boost under BSD, but just
1955 a small improvement under Linux. However, it saves one field in the frame
1956 in all cases.
1957
1958 6. Added more features from the forthcoming Perl 5.10:
1959
1960 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1961 recursion call. It refers to the nth most recently opened parentheses.
1962
1963 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1964 to be opened parentheses.
1965
1966 (c) Conditions that refer to capturing parentheses can be specified
1967 relatively, for example, (?(-2)... or (?(+3)...
1968
1969 (d) \K resets the start of the current match so that everything before
1970 is not part of it.
1971
1972 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1973
1974 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1975 reference syntax.
1976
1977 (g) (?| introduces a group in which the numbering of parentheses in each
1978 alternative starts with the same number.
1979
1980 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1981
1982 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1983 PCRE_INFO_JCHANGED.
1984
1985 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1986 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1987 for detecting groups that can match an empty string.
1988
1989 9. A pattern with a very large number of alternatives (more than several
1990 hundred) was running out of internal workspace during the pre-compile
1991 phase, where pcre_compile() figures out how much memory will be needed. A
1992 bit of new cunning has reduced the workspace needed for groups with
1993 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1994 workspace instead of running out of the 4096 that are available.
1995
1996 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1997
1998 11. Applied patch from Google to remove an optimization that didn't quite work.
1999 The report of the bug said:
2000
2001 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2002 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2003 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2004
2005 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2006 it matched the wrong number of bytes.
2007
2008
2009 Version 7.1 24-Apr-07
2010 ---------------------
2011
2012 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2013 that is more "standard", making use of automake and other Autotools. There
2014 is some re-arrangement of the files and adjustment of comments consequent
2015 on this.
2016
2017 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2018 for recursive directory scanning broke on some systems because the files
2019 are not scanned in any specific order and on different systems the order
2020 was different. A call to "sort" has been inserted into RunGrepTest for the
2021 approprate test as a short-term fix. In the longer term there may be an
2022 alternative.
2023
2024 3. I had an email from Eric Raymond about problems translating some of PCRE's
2025 man pages to HTML (despite the fact that I distribute HTML pages, some
2026 people do their own conversions for various reasons). The problems
2027 concerned the use of low-level troff macros .br and .in. I have therefore
2028 removed all such uses from the man pages (some were redundant, some could
2029 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2030 HTML has been updated to handle .nf/.fi and to complain if it encounters
2031 .br or .in.
2032
2033 4. Updated comments in configure.ac that get placed in config.h.in and also
2034 arranged for config.h to be included in the distribution, with the name
2035 config.h.generic, for the benefit of those who have to compile without
2036 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2037
2038 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2039 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2040 makevp.bat for the current PCRE, using the additional files
2041 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2042
2043 6. A Windows user reported a minor discrepancy with test 2, which turned out
2044 to be caused by a trailing space on an input line that had got lost in his
2045 copy. The trailing space was an accident, so I've just removed it.
2046
2047 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2048 that is needed.
2049
2050 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2051 as "const" (a) because they are and (b) because it helps the PHP
2052 maintainers who have recently made a script to detect big data structures
2053 in the php code that should be moved to the .rodata section. I remembered
2054 to update Builducptable as well, so it won't revert if ucptable.h is ever
2055 re-created.
2056
2057 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2058 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2059 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2060 support is not required. This saves 1.5-2K of code, which is important in
2061 some applications.
2062
2063 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2064 so as not to refer to the tables, even though these functions will never be
2065 called when UTF-8 support is disabled. Otherwise there are problems with a
2066 shared library.
2067
2068 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2069
2070 (a) It was defining its arguments as char * instead of void *.
2071
2072 (b) It was assuming that all moves were upwards in memory; this was true
2073 a long time ago when I wrote it, but is no longer the case.
2074
2075 The emulated memove() is provided for those environments that have neither
2076 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2077 is clearly not the case, as these two bugs were recently reported.
2078
2079 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2080 and Detrail to create the HTML documentation, the .txt form of the man
2081 pages, and it removes trailing spaces from listed files. It also creates
2082 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2083 case, it wraps all the #defines with #ifndefs. This script should be run
2084 before "make dist".
2085
2086 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2087 with Unicode property support.
2088
2089 (a) For a maximizing quantifier, if the two different cases of the
2090 character were of different lengths in their UTF-8 codings (there are
2091 some cases like this - I found 11), and the matching function had to
2092 back up over a mixture of the two cases, it incorrectly assumed they
2093 were both the same length.
2094
2095 (b) When PCRE was configured to use the heap rather than the stack for
2096 recursion during matching, it was not correctly preserving the data for
2097 the other case of a UTF-8 character when checking ahead for a match
2098 while processing a minimizing repeat. If the check also involved
2099 matching a wide character, but failed, corruption could cause an
2100 erroneous result when trying to check for a repeat of the original
2101 character.
2102
2103 13. Some tidying changes to the testing mechanism:
2104
2105 (a) The RunTest script now detects the internal link size and whether there
2106 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2107 values substituted by "configure". (The RunGrepTest script already did
2108 this for UTF-8.) The configure.ac script no longer substitutes the
2109 relevant variables.
2110
2111 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2112 with length and offset values. This means that the output is different
2113 for different internal link sizes. Test 2 is skipped for link sizes
2114 other than 2 because of this, bypassing the problem. Unfortunately,
2115 there was also a test in test 3 (the locale tests) that used /B and
2116 failed for link sizes other than 2. Rather than cut the whole test out,
2117 I have added a new /Z option to pcretest that replaces the length and
2118 offset values with spaces. This is now used to make test 3 independent
2119 of link size. (Test 2 will be tidied up later.)
2120
2121 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2122 segmentation fault instead of returning the appropriate error message.
2123
2124 15. In multiline mode when the newline sequence was set to "any", the pattern
2125 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2126 This doesn't seem right; it now treats the CRLF combination as the line
2127 ending, and so does not match in that case. It's only a pattern such as ^$
2128 that would hit this one: something like ^ABC$ would have failed after \r
2129 and then tried again after \r\n.
2130
2131 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2132 in an attempt to make files that differ only in their line terminators
2133 compare equal. This works on Linux.
2134
2135 17. Under certain error circumstances pcregrep might try to free random memory
2136 as it exited. This is now fixed, thanks to valgrind.
2137
2138 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2139 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2140 was because its rules for how to advance for /g after matching an empty
2141 string at the end of a line did not allow for this case. They now check for
2142 it specially.
2143
2144 20. pcretest is supposed to handle patterns and data of any length, by
2145 extending its buffers when necessary. It was getting this wrong when the
2146 buffer for a data line had to be extended.
2147
2148 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2149 CRLF as a newline sequence.
2150
2151 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2152 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2153 I have nevertheless tidied it up.
2154
2155 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2156
2157 24. Added a man page for pcre-config.
2158
2159
2160 Version 7.0 19-Dec-06
2161 ---------------------
2162
2163 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2164 moving to gcc 4.1.1.
2165
2166 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2167 sys/time.h, which is documented as needed for this function. It doesn't
2168 seem to matter on Linux, but it showed up on some releases of OS X.
2169
2170 3. It seems that there are systems where bytes whose values are greater than
2171 127 match isprint() in the "C" locale. The "C" locale should be the
2172 default when a C program starts up. In most systems, only ASCII printing
2173 characters match isprint(). This difference caused the output from pcretest
2174 to vary, making some of the tests fail. I have changed pcretest so that:
2175
2176 (a) When it is outputting text in the compiled version of a pattern, bytes
2177 other than 32-126 are always shown as hex escapes.
2178
2179 (b) When it is outputting text that is a matched part of a subject string,
2180 it does the same, unless a different locale has been set for the match
2181 (using the /L modifier). In this case, it uses isprint() to decide.
2182
2183 4. Fixed a major bug that caused incorrect computation of the amount of memory
2184 required for a compiled pattern when options that changed within the
2185 pattern affected the logic of the preliminary scan that determines the
2186 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2187 that the computed length was too small. The symptoms of this bug were
2188 either the PCRE error "internal error: code overflow" from pcre_compile(),
2189 or a glibc crash with a message such as "pcretest: free(): invalid next
2190 size (fast)". Examples of patterns that provoked this bug (shown in
2191 pcretest format) are:
2192
2193 /(?-x: )/x
2194 /(?x)(?-x: \s*#\s*)/
2195 /((?i)[\x{c0}])/8
2196 /(?i:[\x{c0}])/8
2197
2198 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2199 is now done differently.
2200
2201 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2202 wrapper classes; (b) implement a new function in the C++ scanner that is
2203 more efficient than the old way of doing things because it avoids levels of
2204 recursion in the regex matching; (c) add a paragraph to the documentation
2205 for the FullMatch() function.
2206
2207 6. The escape sequence \n was being treated as whatever was defined as
2208 "newline". Not only was this contrary to the documentation, which states
2209 that \n is character 10 (hex 0A), but it also went horribly wrong when
2210 "newline" was defined as CRLF. This has been fixed.
2211
2212 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2213 was being set to -1 for the "end of line" case (supposedly a value that no
2214 character can have). Though this value is never used (the check for end of
2215 line is "zero bytes in current character"), it caused compiler complaints.
2216 I've changed it to 0xffffffff.
2217
2218 8. In pcre_version.c, the version string was being built by a sequence of
2219 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2220 string (as it is for production releases) called a macro with an empty
2221 argument. The C standard says the result of this is undefined. The gcc
2222 compiler treats it as an empty string (which was what was wanted) but it is
2223 reported that Visual C gives an error. The source has been hacked around to
2224 avoid this problem.
2225
2226 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2227 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2228 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2229 of them did).
2230
2231 10. Originally, pcretest opened its input and output without "b"; then I was
2232 told that "b" was needed in some environments, so it was added for release
2233 5.0 to both the input and output. (It makes no difference on Unix-like
2234 systems.) Later I was told that it is wrong for the input on Windows. I've
2235 now abstracted the modes into two macros, to make it easier to fiddle with
2236 them, and removed "b" from the input mode under Windows.
2237
2238 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2239
2240 12. Added -help and --help to pcretest as an official way of being reminded
2241 of the options.
2242
2243 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2244 and pcrecpp.cc because they annoy compilers at high warning levels.
2245
2246 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2247
2248 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2249 scripts are not C programs :-) and which was not noticed because it works
2250 on Linux.
2251
2252 16. pcretest is supposed to handle any length of pattern and data line (as one
2253 line or as a continued sequence of lines) by extending its input buffer if
2254 necessary. This feature was broken for very long pattern lines, leading to
2255 a string of junk being passed to pcre_compile() if the pattern was longer
2256 than about 50K.
2257
2258 17. I have done a major re-factoring of the way pcre_compile() computes the
2259 amount of memory needed for a compiled pattern. Previously, there was code
2260 that made a preliminary scan of the pattern in order to do this. That was
2261 OK when PCRE was new, but as the facilities have expanded, it has become
2262 harder and harder to keep it in step with the real compile phase, and there
2263 have been a number of bugs (see for example, 4 above). I have now found a
2264 cunning way of running the real compile function in a "fake" mode that
2265 enables it to compute how much memory it would need, while actually only
2266 ever using a few hundred bytes of working memory and without too many
2267 tests of the mode. This should make future maintenance and development
2268 easier. A side effect of this work is that the limit of 200 on the nesting
2269 depth of parentheses has been removed (though this was never a serious
2270 limitation, I suspect). However, there is a downside: pcre_compile() now
2271 runs more slowly than before (30% or more, depending on the pattern). I
2272 hope this isn't a big issue. There is no effect on runtime performance.
2273
2274 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2275 newline (only possible for the last line of a file) and it was a
2276 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2277
2278 19. Added additional timing features to pcretest. (1) The -tm option now times
2279 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2280 separate command line item, by a number that specifies the number of
2281 repeats to use when timing. The default is 50000; this gives better
2282 precision, but takes uncomfortably long for very large patterns.
2283
2284 20. Extended pcre_study() to be more clever in cases where a branch of a
2285 subpattern has no definite first character. For example, (a*|b*)[cd] would
2286 previously give no result from pcre_study(). Now it recognizes that the
2287 first character must be a, b, c, or d.
2288
2289 21. There was an incorrect error "recursive call could loop indefinitely" if
2290 a subpattern (or the entire pattern) that was being tested for matching an
2291 empty string contained only one non-empty item after a nested subpattern.
2292 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2293 incorrectly, because the \d was being skipped in the check.
2294
2295 22. The pcretest program now has a new pattern option /B and a command line
2296 option -b, which is equivalent to adding /B to every pattern. This causes
2297 it to show the compiled bytecode, without the additional information that
2298 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2299 is the same as /B/I).
2300
2301 23. A new optimization is now able automatically to treat some sequences such
2302 as a*b as a*+b. More specifically, if something simple (such as a character
2303 or a simple class like \d) has an unlimited quantifier, and is followed by
2304 something that cannot possibly match the quantified thing, the quantifier
2305 is automatically "possessified".
2306
2307 24. A recursive reference to a subpattern whose number was greater than 39
2308 went wrong under certain circumstances in UTF-8 mode. This bug could also
2309 have affected the operation of pcre_study().
2310
2311 25. Realized that a little bit of performance could be had by replacing
2312 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2313
2314 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2315
2316 27. Possessive quantifiers such as a++ were previously implemented by turning
2317 them into atomic groups such as ($>a+). Now they have their own opcodes,
2318 which improves performance. This includes the automatically created ones
2319 from 23 above.
2320
2321 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2322 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2323 the first matched character to be a colon. This applied both to named and
2324 numbered groups.
2325
2326 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2327
2328 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2329 building PCRE on Windows easier, so I have included it in the distribution.
2330
2331 31. There is now a check in pcretest against a ridiculously large number being
2332 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2333 loop, the loop is abandoned.
2334
2335 32. Forward references to subpatterns in conditions such as (?(2)...) where
2336 subpattern 2 is defined later cause pcre_compile() to search forwards in
2337 the pattern for the relevant set of parentheses. This search went wrong
2338 when there were unescaped parentheses in a character class, parentheses
2339 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2340
2341 33. "Subroutine" calls and backreferences were previously restricted to
2342 referencing subpatterns earlier in the regex. This restriction has now
2343 been removed.
2344
2345 34. Added a number of extra features that are going to be in Perl 5.10. On the
2346 whole, these are just syntactic alternatives for features that PCRE had
2347 previously implemented using the Python syntax or my own invention. The
2348 other formats are all retained for compatibility.
2349
2350 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2351 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2352 also .NET compatible.
2353
2354 (b) A recursion or subroutine call to a named group can now be defined as
2355 (?&name) as well as (?P>name).
2356
2357 (c) A backreference to a named group can now be defined as \k<name> or
2358 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2359 5.10, are also .NET compatible.
2360
2361 (d) A conditional reference to a named group can now use the syntax
2362 (?(<name>) or (?('name') as well as (?(name).
2363
2364 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2365 groups (named and numbered) that are never evaluated inline, but can be
2366 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2367 is always false. There may be only one alternative in such a group.
2368
2369 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2370 as the simple (?(R). The condition is true only if the most recent
2371 recursion is that of the given number or name. It does not search out
2372 through the entire recursion stack.
2373
2374 (g) The escape \gN or \g{N} has been added, where N is a positive or
2375 negative number, specifying an absolute or relative reference.
2376
2377 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2378 some "unreachable code" warnings.
2379
2380 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2381 things, this adds five new scripts.
2382
2383 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2384 There were also incompatibilities regarding the handling of \Q..\E inside
2385 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2386 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2387
2388 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2389 matches an empty string, and forcibly breaks the loop. There were bugs in
2390 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2391 against aaaa the result was just "a" rather than "aaaa", for example. Two
2392 separate and independent bugs (that affected different cases) have been
2393 fixed.
2394
2395 39. Refactored the code to abolish the use of different opcodes for small
2396 capturing bracket numbers. This is a tidy that I avoided doing when I
2397 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2398 The new approach is not only tidier, it makes it possible to reduce the
2399 memory needed to fix the previous bug (38).
2400
2401 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2402 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2403 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2404 mode.
2405
2406 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2407 report.
2408
2409 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2410 copy construction and assignment in the C++ wrapper.
2411
2412 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2413 couple of bugs that could have given wrong results in the "--newline=crlf"
2414 case.
2415
2416 44. Added a number of casts and did some reorganization of signed/unsigned int
2417 variables following suggestions from Dair Grant. Also renamed the variable
2418 "this" as "item" because it is a C++ keyword.
2419
2420 45. Arranged for dftables to add
2421
2422 #include "pcre_internal.h"
2423
2424 to pcre_chartables.c because without it, gcc 4.x may remove the array
2425 definition from the final binary if PCRE is built into a static library and
2426 dead code stripping is activated.
2427
2428 46. For an unanchored pattern, if a match attempt fails at the start of a
2429 newline sequence, and the newline setting is CRLF or ANY, and the next two
2430 characters are CRLF, advance by two characters instead of one.
2431
2432
2433 Version 6.7 04-Jul-06
2434 ---------------------
2435
2436 1. In order to handle tests when input lines are enormously long, pcretest has
2437 been re-factored so that it automatically extends its buffers when
2438 necessary. The code is crude, but this _is_ just a test program. The
2439 default size has been increased from 32K to 50K.
2440
2441 2. The code in pcre_study() was using the value of the re argument before
2442 testing it for NULL. (Of course, in any sensible call of the function, it
2443 won't be NULL.)
2444
2445 3. The memmove() emulation function in pcre_internal.h, which is used on
2446 systems that lack both memmove() and bcopy() - that is, hardly ever -
2447 was missing a "static" storage class specifier.
2448
2449 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2450 containing an extended class (one that cannot be represented by a bitmap
2451 because it contains high-valued characters or Unicode property items, e.g.
2452 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2453 pattern, but PCRE should not loop if you do not (it no longer does).
2454 [Detail: two cases were found: (a) a repeated subpattern containing an
2455 extended class; (b) a recursive reference to a subpattern that followed a
2456 previous extended class. It wasn't skipping over the extended class
2457 correctly when UTF-8 mode was not set.]
2458
2459 5. A negated single-character class was not being recognized as fixed-length
2460 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2461 compile error "lookbehind assertion is not fixed length".
2462
2463 6. The RunPerlTest auxiliary script was showing an unexpected difference
2464 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2465 write a Perl script that can interpret lines of an input file either as
2466 byte characters or as UTF-8, which is what "perltest" was being required to
2467 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2468 can't do is switch easily at run time between having the "use utf8;" pragma
2469 or not. In the end, I fudged it by using the RunPerlTest script to insert
2470 "use utf8;" explicitly for the UTF-8 tests.
2471
2472 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2473 the end of the subject string, contrary to the documentation and to what
2474 Perl does. This was true of both matching functions. Now it matches only at
2475 the start of the subject and immediately after *internal* newlines.
2476
2477 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2478 a pointer to an int instead of a pointer to an unsigned long int. This
2479 caused problems on 64-bit systems.
2480
2481 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2482 instance of the 'standard' template library not being so standard".
2483
2484 10. There was no check on the number of named subpatterns nor the maximum
2485 length of a subpattern name. The product of these values is used to compute
2486 the size of the memory block for a compiled pattern. By supplying a very
2487 long subpattern name and a large number of named subpatterns, the size
2488 computation could be caused to overflow. This is now prevented by limiting
2489 the length of names to 32 characters, and the number of named subpatterns
2490 to 10,000.
2491
2492 11. Subpatterns that are repeated with specific counts have to be replicated in
2493 the compiled pattern. The size of memory for this was computed from the
2494 length of the subpattern and the repeat count. The latter is limited to
2495 65535, but there was no limit on the former, meaning that integer overflow
2496 could in principle occur. The compiled length of a repeated subpattern is
2497 now limited to 30,000 bytes in order to prevent this.
2498
2499 12. Added the optional facility to have named substrings with the same name.
2500
2501 13. Added the ability to use a named substring as a condition, using the
2502 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2503 are numbers (not recommended). Forward references are permitted.
2504
2505 14. Added forward references in named backreferences (if you see what I mean).
2506
2507 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2508 pattern could run off the end of the subject. For example, the pattern
2509 "(?s)(.{1,5})"8 did this with the subject "ab".
2510
2511 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2512 PCRE_CASELESS was set when matching characters that were quantified with ?
2513 or *.
2514
2515 17. A character class other than a single negated character that had a minimum
2516 but no maximum quantifier - for example [ab]{6,} - was not handled
2517 correctly by pce_dfa_exec(). It would match only one character.
2518
2519 18. A valid (though odd) pattern that looked like a POSIX character
2520 class but used an invalid character after [ (for example [[,abc,]]) caused
2521 pcre_compile() to give the error "Failed: internal error: code overflow" or
2522 in some cases to crash with a glibc free() error. This could even happen if
2523 the pattern terminated after [[ but there just happened to be a sequence of
2524 letters, a binary zero, and a closing ] in the memory that followed.
2525
2526 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2527 over the years. Originally (before any Unicode support), just the bottom 8
2528 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2529 output from "man perlunicode" includes this:
2530
2531 The regular expression compiler produces polymorphic opcodes. That
2532 is, the pattern adapts to the data and automatically switches to
2533 the Unicode character scheme when presented with Unicode data--or
2534 instead uses a traditional byte scheme when presented with byte
2535 data.
2536
2537 Sadly, a wide octal escape does not cause a switch, and in a string with
2538 no other multibyte characters, these octal escapes are treated as before.
2539 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2540 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2541 Unicode string.
2542
2543 I have not perpetrated such confusion in PCRE. Up till now, it took just
2544 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2545 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2546 translate to the appropriate multibyte character.
2547
2548 29. Applied some refactoring to reduce the number of warnings from Microsoft
2549 and Borland compilers. This has included removing the fudge introduced
2550 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2551 a warning about an unused variable.
2552
2553 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2554 characters since release 4.0, because Perl (from release 5.004) does not.
2555 [Or at least, is documented not to: some releases seem to be in conflict
2556 with the documentation.] However, when a pattern was studied with
2557 pcre_study() and all its branches started with \s, PCRE still included VT
2558 as a possible starting character. Of course, this did no harm; it just
2559 caused an unnecessary match attempt.
2560
2561 22. Removed a now-redundant internal flag bit that recorded the fact that case
2562 dependency changed within the pattern. This was once needed for "required
2563 byte" processing, but is no longer used. This recovers a now-scarce options
2564 bit. Also moved the least significant internal flag bit to the most-
2565 significant bit of the word, which was not previously used (hangover from
2566 the days when it was an int rather than a uint) to free up another bit for
2567 the future.
2568
2569 23. Added support for CRLF line endings as well as CR and LF. As well as the
2570 default being selectable at build time, it can now be changed at runtime
2571 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2572 specify that it is scanning data with non-default line endings.
2573
2574 24. Changed the definition of CXXLINK to make it agree with the definition of
2575 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2576
2577 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2578 recursions. This makes a big different to stack usage for some patterns.
2579
2580 26. If a subpattern containing a named recursion or subroutine reference such
2581 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2582 the space required for the compiled pattern went wrong and gave too small a
2583 value. Depending on the environment, this could lead to "Failed: internal
2584 error: code overflow at offset 49" or "glibc detected double free or
2585 corruption" errors.
2586
2587 27. Applied patches from Google (a) to support the new newline modes and (b) to
2588 advance over multibyte UTF-8 characters in GlobalReplace.
2589
2590 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2591 difference for some implementation of PCRE in some Windows version.
2592
2593 29. Added some extra testing facilities to pcretest:
2594
2595 \q<number> in a data line sets the "match limit" value
2596 \Q<number> in a data line sets the "match recursion limt" value
2597 -S <number> sets the stack size, where <number> is in megabytes
2598
2599 The -S option isn't available for Windows.
2600
2601
2602 Version 6.6 06-Feb-06
2603 ---------------------
2604
2605 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2606 in pcreposix.h. I have copied the definition from pcre.h.
2607
2608 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2609 because pcre.h is no longer a built file.
2610
2611 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2612 not normally included in the compiled code.
2613
2614
2615 Version 6.5 01-Feb-06
2616 ---------------------
2617
2618 1. When using the partial match feature with pcre_dfa_exec(), it was not
2619 anchoring the second and subsequent partial matches at the new starting
2620 point. This could lead to incorrect results. For example, with the pattern
2621 /1234/, partially matching against "123" and then "a4" gave a match.
2622
2623 2. Changes to pcregrep:
2624
2625 (a) All non-match returns from pcre_exec() were being treated as failures
2626 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2627 error message is output. Some extra information is given for the
2628 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2629 probably the only errors that are likely to be caused by users (by
2630 specifying a regex that has nested indefinite repeats, for instance).
2631 If there are more than 20 of these errors, pcregrep is abandoned.
2632
2633 (b) A binary zero was treated as data while matching, but terminated the
2634 output line if it was written out. This has been fixed: binary zeroes
2635 are now no different to any other data bytes.
2636
2637 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2638 used to set a locale for matching. The --locale=xxxx long option has
2639 been added (no short equivalent) to specify a locale explicitly on the
2640 pcregrep command, overriding the environment variables.
2641
2642 (d) When -B was used with -n, some line numbers in the output were one less
2643 than they should have been.
2644
2645 (e) Added the -o (--only-matching) option.
2646
2647 (f) If -A or -C was used with -c (count only), some lines of context were
2648 accidentally printed for the final match.
2649
2650 (g) Added the -H (--with-filename) option.
2651
2652 (h) The combination of options -rh failed to suppress file names for files
2653 that were found from directory arguments.
2654
2655 (i) Added the -D (--devices) and -d (--directories) options.
2656
2657 (j) Added the -F (--fixed-strings) option.
2658
2659 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2660
2661 (l) Added the --colo(u)r option.
2662
2663 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2664 is not present by default.
2665
2666 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2667 items such as (?R) or (?1), when the recursion could match a number of
2668 alternatives. If it matched one of the alternatives, but subsequently,
2669 outside the recursion, there was a failure, the code tried to back up into
2670 the recursion. However, because of the way PCRE is implemented, this is not
2671 possible, and the result was an incorrect result from the match.
2672
2673 In order to prevent this happening, the specification of recursion has
2674 been changed so that all such subpatterns are automatically treated as
2675 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2676
2677 4. I had overlooked the fact that, in some locales, there are characters for
2678 which isalpha() is true but neither isupper() nor islower() are true. In
2679 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2680 and ordfeminine) are like this. This affected the treatment of \w and \W
2681 when they appeared in character classes, but not when they appeared outside
2682 a character class. The bit map for "word" characters is now created
2683 separately from the results of isalnum() instead of just taking it from the
2684 upper, lower, and digit maps. (Plus the underscore character, of course.)
2685
2686 5. The above bug also affected the handling of POSIX character classes such as
2687 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2688 permanent tables. Instead, the bit maps for such a class were previously
2689 created as the appropriate unions of the upper, lower, and digit bitmaps.
2690 Now they are created by subtraction from the [[:word:]] class, which has
2691 its own bitmap.
2692
2693 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2694 It is created by subtracting the vertical space characters (\x09, \x0a,
2695 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2696 subtraction was done in the overall bitmap for a character class, meaning
2697 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2698 be recognized. This bug has been fixed.
2699
2700 7. Patches from the folks at Google:
2701
2702 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2703 real life, but is still worth protecting against".
2704
2705 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2706 regular expressions".
2707
2708 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2709 have it.
2710
2711 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2712 "configure" and the latter not, in order to fix a problem somebody had
2713 with compiling the Arg class on HP-UX.
2714
2715 (e) Improve the error-handling of the C++ wrapper a little bit.
2716
2717 (f) New tests for checking recursion limiting.
2718
2719 8. The pcre_memmove() function, which is used only if the environment does not
2720 have a standard memmove() function (and is therefore rarely compiled),
2721 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2722 returning a result (though PCRE never actually uses the result).
2723
2724 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2725 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2726 returned instead of calling malloc() with an overflowing number that would
2727 most likely cause subsequent chaos.
2728
2729 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2730
2731 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2732 with this option is matched, the nmatch and pmatch options of regexec() are
2733 ignored.
2734
2735 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2736 provided in case anyone wants to the the POSIX interface with UTF-8
2737 strings.
2738
2739 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2740 C++ linking (needed for some HP-UX environments).
2741
2742 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2743 (unused parameter) and in the pcre_printint() function (omitted "default"
2744 switch label when the default is to do nothing).
2745
2746 15. Added some code to make it possible, when PCRE is compiled as a C++
2747 library, to replace subject pointers for pcre_exec() with a smart pointer
2748 class, thus making it possible to process discontinuous strings.
2749
2750 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2751 much the same function. They were added by different people who were trying
2752 to make PCRE easy to compile on non-Unix systems. It has been suggested
2753 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2754 for compiling on Windows systems. I have therefore replaced it with
2755 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2756 defaults to "extern" for C or "extern C" for C++, which works fine on
2757 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2758 SCOPE with something explicit in config.h. In addition:
2759
2760 (a) pcreposix.h still had just "extern" instead of either of these macros;
2761 I have replaced it with PCRE_DATA_SCOPE.
2762
2763 (b) Functions such as _pcre_xclass(), which are internal to the library,
2764 but external in the C sense, all had PCRE_EXPORT in their definitions.
2765 This is apparently wrong for the Windows case, so I have removed it.
2766 (It makes no difference on Unix-like systems.)
2767
2768 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2769 of recursive calls to match(). This is different to MATCH_LIMIT because
2770 that limits the total number of calls to match(), not all of which increase
2771 the depth of recursion. Limiting the recursion depth limits the amount of
2772 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2773 when PCRE is compiled, and changed at run time. A patch from Google adds
2774 this functionality to the C++ interface.
2775
2776 18. Changes to the handling of Unicode character properties:
2777
2778 (a) Updated the table to Unicode 4.1.0.
2779
2780 (b) Recognize characters that are not in the table as "Cn" (undefined).
2781
2782 (c) I revised the way the table is implemented to a much improved format
2783 which includes recognition of ranges. It now supports the ranges that
2784 are defined in UnicodeData.txt, and it also amalgamates other
2785 characters into ranges. This has reduced the number of entries in the
2786 table from around 16,000 to around 3,000, thus reducing its size
2787 considerably. I realized I did not need to use a tree structure after
2788 all - a binary chop search is just as efficient. Having reduced the
2789 number of entries, I extended their size from 6 bytes to 8 bytes to
2790 allow for more data.
2791
2792 (d) Added support for Unicode script names via properties such as \p{Han}.
2793
2794 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2795 matching that character.
2796
2797 20. When matching a repeated Unicode property with a minimum greater than zero,
2798 (for example \pL{2,}), PCRE could look past the end of the subject if it
2799 reached it while seeking the minimum number of characters. This could
2800 happen only if some of the characters were more than one byte long, because
2801 there is a check for at least the minimum number of bytes.
2802
2803 21. Refactored the implementation of \p and \P so as to be more general, to
2804 allow for more different types of property in future. This has changed the
2805 compiled form incompatibly. Anybody with saved compiled patterns that use
2806 \p or \P will have to recompile them.
2807
2808 22. Added "Any" and "L&" to the supported property types.
2809
2810 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2811 but give a compile time error if the value is greater than 0xff.
2812
2813 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2814 accidentally not being installed or uninstalled.
2815
2816 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2817 made were to insert the current release number. This seemed silly, because
2818 it made things harder for people building PCRE on systems that don't run
2819 "configure". I have turned pcre.h into a distributed file, no longer built
2820 by "configure", with the version identification directly included. There is
2821 no longer a pcre.h.in file.
2822
2823 However, this change necessitated a change to the pcre-config script as
2824 well. It is built from pcre-config.in, and one of the substitutions was the
2825 release number. I have updated configure.ac so that ./configure now finds
2826 the release number by grepping pcre.h.
2827
2828 26. Added the ability to run the tests under valgrind.
2829
2830
2831 Version 6.4 05-Sep-05
2832 ---------------------
2833
2834 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2835 "--" to be printed when multiple files were scanned, even when none of the
2836 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2837 consider it to be a bug, and have restored the previous behaviour.
2838
2839 2. A couple of code tidies to get rid of compiler warnings.
2840
2841 3. The pcretest program used to cheat by referring to symbols in the library
2842 whose names begin with _pcre_. These are internal symbols that are not
2843 really supposed to be visible externally, and in some environments it is
2844 possible to suppress them. The cheating is now confined to including
2845 certain files from the library's source, which is a bit cleaner.
2846
2847 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2848 file's purpose clearer.
2849
2850 5. Reorganized pcre_ucp_findchar().
2851
2852
2853 Version 6.3 15-Aug-05
2854 ---------------------
2855
2856 1. The file libpcre.pc.in did not have general read permission in the tarball.
2857
2858 2. There were some problems when building without C++ support:
2859
2860 (a) If C++ support was not built, "make install" and "make test" still
2861 tried to test it.
2862
2863 (b) There were problems when the value of CXX was explicitly set. Some
2864 changes have been made to try to fix these, and ...
2865
2866 (c) --disable-cpp can now be used to explicitly disable C++ support.
2867
2868 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2869 backslash in a target when C++ was disabled. This confuses some
2870 versions of "make", apparently. Using an intermediate variable solves
2871 this. (Same for CPP_LOBJ.)
2872
2873 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2874 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2875 necessary on certain architectures.
2876
2877 4. Added a setting of -export-symbols-regex to the link command to remove
2878 those symbols that are exported in the C sense, but actually are local
2879 within the library, and not documented. Their names all begin with
2880 "_pcre_". This is not a perfect job, because (a) we have to except some
2881 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2882 available (and never for static libraries). I have made a note to try to
2883 find a way round (a) in the future.
2884
2885
2886 Version 6.2 01-Aug-05
2887 ---------------------
2888
2889 1. There was no test for integer overflow of quantifier values. A construction
2890 such as {1111111111111111} would give undefined results. What is worse, if
2891 a minimum quantifier for a parenthesized subpattern overflowed and became
2892 negative, the calculation of the memory size went wrong. This could have
2893 led to memory overwriting.
2894
2895 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2896
2897 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2898 operating environments where this matters.
2899
2900 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2901 PCRE options from within the C++ wrapper.
2902
2903 5. Named capturing subpatterns were not being correctly counted when a pattern
2904 was compiled. This caused two problems: (a) If there were more than 100
2905 such subpatterns, the calculation of the memory needed for the whole
2906 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2907 back references of the form \12, where the number was greater than 9, were
2908 not recognized as back references, even though there were sufficient
2909 previous subpatterns.
2910
2911 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2912 versions of gcc, e.g. 2.95.4.
2913
2914
2915 Version 6.1 21-Jun-05
2916 ---------------------
2917
2918 1. There was one reference to the variable "posix" in pcretest.c that was not
2919 surrounded by "#if !defined NOPOSIX".
2920
2921 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2922 the cross-check on the old pcre_info() function, for the benefit of the
2923 cut-down version of PCRE that is currently imported into Exim.
2924
2925 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2926 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2927 patterns that start (?i) but I don't think that matters. The use of (?i) is
2928 just an example; this all applies to the other options as well.
2929
2930 4. Since libtool seems to echo the compile commands it is issuing, the output
2931 from "make" can be reduced a bit by putting "@" in front of each libtool
2932 compile command.
2933
2934 5. Patch from the folks at Google for configure.in to be a bit more thorough
2935 in checking for a suitable C++ installation before trying to compile the
2936 C++ stuff. This should fix a reported problem when a compiler was present,
2937 but no suitable headers.
2938
2939 6. The man pages all had just "PCRE" as their title. I have changed them to
2940 be the relevant file name. I have also arranged that these names are
2941 retained in the file doc/pcre.txt, which is a concatenation in text format
2942 of all the man pages except the little individual ones for each function.
2943
2944 7. The NON-UNIX-USE file had not been updated for the different set of source
2945 files that come with release 6. I also added a few comments about the C++
2946 wrapper.
2947
2948
2949 Version 6.0 07-Jun-05
2950 ---------------------
2951
2952 1. Some minor internal re-organization to help with my DFA experiments.
2953
2954 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2955 didn't matter for the library itself when fully configured, but did matter
2956 when compiling without UCP support, or within Exim, where the ucp files are
2957 not imported.
2958
2959 3. Refactoring of the library code to split up the various functions into
2960 different source modules. The addition of the new DFA matching code (see
2961 below) to a single monolithic source would have made it really too
2962 unwieldy, quite apart from causing all the code to be include in a
2963 statically linked application, when only some functions are used. This is
2964 relevant even without the DFA addition now that patterns can be compiled in
2965 one application and matched in another.
2966
2967 The downside of splitting up is that there have to be some external
2968 functions and data tables that are used internally in different modules of
2969 the library but which are not part of the API. These have all had their
2970 names changed to start with "_pcre_" so that they are unlikely to clash
2971 with other external names.
2972
2973 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2974 a different (DFA) algorithm. Although it is slower than the original
2975 function, it does have some advantages for certain types of matching
2976 problem.
2977
2978 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2979 including restarting after a partial match.
2980
2981 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2982 defined when compiling for Windows was sent to me. I have put it into the
2983 code, though I have no means of testing or verifying it.
2984
2985 7. Added the pcre_refcount() auxiliary function.
2986
2987 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2988 match before or at the first newline in the subject string. In pcretest,
2989 the /f option on a pattern can be used to set this.
2990
2991 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2992 would behave wrongly. This has been present in PCRE since release 4.0.
2993
2994 10. A number of changes to the pcregrep command:
2995
2996 (a) Refactored how -x works; insert ^(...)$ instead of setting
2997 PCRE_ANCHORED and checking the length, in preparation for adding
2998 something similar for -w.
2999
3000 (b) Added the -w (match as a word) option.
3001
3002 (c) Refactored the way lines are read and buffered so as to have more
3003 than one at a time available.
3004
3005 (d) Implemented a pcregrep test script.
3006
3007 (e) Added the -M (multiline match) option. This allows patterns to match
3008 over several lines of the subject. The buffering ensures that at least
3009 8K, or the rest of the document (whichever is the shorter) is available
3010 for matching (and similarly the previous 8K for lookbehind assertions).
3011
3012 (f) Changed the --help output so that it now says
3013
3014 -w, --word-regex(p)
3015
3016 instead of two lines, one with "regex" and the other with "regexp"
3017 because that confused at least one person since the short forms are the
3018 same. (This required a bit of code, as the output is generated
3019 automatically from a table. It wasn't just a text change.)
3020
3021 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3022 option but starts with a hyphen. Could be a pattern or a path name
3023 starting with a hyphen, for instance.
3024
3025 (h) "-" can be given as a file name to represent stdin.
3026
3027 (i) When file names are being printed, "(standard input)" is used for
3028 the standard input, for compatibility with GNU grep. Previously
3029 "<stdin>" was used.
3030
3031 (j) The option --label=xxx can be used to supply a name to be used for
3032 stdin when file names are being printed. There is no short form.
3033
3034 (k) Re-factored the options decoding logic because we are going to add
3035 two more options that take data. Such options can now be given in four
3036 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3037
3038 (l) Added the -A, -B, and -C options for requesting that lines of context
3039 around matches be printed.
3040
3041 (m) Added the -L option to print the names of files that do not contain
3042 any matching lines, that is, the complement of -l.
3043
3044 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3045 continue to scan other files.
3046
3047 (o) The -s option was incorrectly implemented. For compatibility with other
3048 greps, it now suppresses the error message for a non-existent or non-
3049 accessible file (but not the return code). There is a new option called
3050 -q that suppresses the output of matching lines, which was what -s was
3051 previously doing.
3052
3053 (p) Added --include and --exclude options to specify files for inclusion
3054 and exclusion when recursing.
3055
3056 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3057 Hopefully, it now does.
3058
3059 12. Missing cast in pcre_study().
3060
3061 13. Added an "uninstall" target to the makefile.
3062
3063 14. Replaced "extern" in the function prototypes in Makefile.in with
3064 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3065 world, but is set differently for Windows.
3066
3067 15. Added a second compiling function called pcre_compile2(). The only
3068 difference is that it has an extra argument, which is a pointer to an
3069 integer error code. When there is a compile-time failure, this is set
3070 non-zero, in addition to the error test pointer being set to point to an
3071 error message. The new argument may be NULL if no error number is required
3072 (but then you may as well call pcre_compile(), which is now just a
3073 wrapper). This facility is provided because some applications need a
3074 numeric error indication, but it has also enabled me to tidy up the way
3075 compile-time errors are handled in the POSIX wrapper.
3076
3077 16. Added VPATH=.libs to the makefile; this should help when building with one
3078 prefix path and installing with another. (Or so I'm told by someone who
3079 knows more about this stuff than I do.)
3080
3081 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3082 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3083 match everything, including newlines. This is not POSIX-compatible, but
3084 somebody wanted the feature. From pcretest it can be activated by using
3085 both the P and the s flags.
3086
3087 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3088
3089 19. libpcre.pc was being incorrectly installed as executable.
3090
3091 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3092 it now also looks for '\r' so that it will work unmodified on Windows.
3093
3094 21. Added Google's contributed C++ wrapper to the distribution.
3095
3096 22. Added some untidy missing memory free() calls in pcretest, to keep
3097 Electric Fence happy when testing.
3098
3099
3100
3101 Version 5.0 13-Sep-04
3102 ---------------------
3103
3104 1. Internal change: literal characters are no longer packed up into items
3105 containing multiple characters in a single byte-string. Each character
3106 is now matched using a separate opcode. However, there may be more than one
3107 byte in the character in UTF-8 mode.
3108
3109 2. The pcre_callout_block structure has two new fields: pattern_position and
3110 next_item_length. These contain the offset in the pattern to the next match
3111 item, and its length, respectively.
3112
3113 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3114 insertion of callouts before each pattern item. Added the /C option to
3115 pcretest to make use of this.
3116
3117 4. On the advice of a Windows user, the lines
3118
3119 #if defined(_WIN32) || defined(WIN32)
3120 _setmode( _fileno( stdout ), 0x8000 );
3121 #endif /* defined(_WIN32) || defined(WIN32) */
3122
3123 have been added to the source of pcretest. This apparently does useful
3124 magic in relation to line terminators.
3125
3126 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3127 for the benefit of those environments where the "b" makes a difference.
3128
3129 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3130 to know about it. I have put a hack into configure.in that adds in code
3131 to set GCC=yes if CC=icc. This seems to end up at a point in the
3132 generated configure script that is early enough to affect the setting of
3133 compiler options, which is what is needed, but I have no means of testing
3134 whether it really works. (The user who reported this had patched the
3135 generated configure script, which of course I cannot do.)
3136
3137 LATER: After change 22 below (new libtool files), the configure script
3138 seems to know about icc (and also ecc). Therefore, I have commented out
3139 this hack in configure.in.
3140
3141 7. Added support for pkg-config (2 patches were sent in).
3142
3143 8. Negated POSIX character classes that used a combination of internal tables
3144 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3145 [[:^ascii]]. Typically, they would match almost any characters. The other
3146 POSIX classes were not broken in this way.
3147
3148 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3149 to find the match, as PCRE was deluded into thinking that the match had to
3150 start at the start point or following a newline. The same bug applied to
3151 patterns with negative forward assertions or any backward assertions
3152 preceding ".*" at the start, unless the pattern required a fixed first
3153 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3154
3155 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3156 starting at the last subject character, bytes beyond the end of the subject
3157 string were read.
3158
3159 11. Renamed the variable "class" as "classbits" to make life easier for C++
3160 users. (Previously there was a macro definition, but it apparently wasn't
3161 enough.)
3162
3163 12. Added the new field "tables" to the extra data so that tables can be passed
3164 in at exec time, or the internal tables can be re-selected. This allows
3165 a compiled regex to be saved and re-used at a later time by a different
3166 program that might have everything at different addresses.
3167
3168 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3169 -R library as well as a -L library.
3170
3171 14. The debugging options of pcretest (-d on the command line or D on a
3172 pattern) showed incorrect output for anything following an extended class
3173 that contained multibyte characters and which was followed by a quantifier.
3174
3175 15. Added optional support for general category Unicode character properties
3176 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3177 support. It adds about 90K to the size of the library. The meanings of the
3178 inbuilt class escapes such as \d and \s have NOT been changed.
3179
3180 16. Updated pcredemo.c to include calls to free() to release the memory for the
3181 compiled pattern.
3182
3183 17. The generated file chartables.c was being created in the source directory
3184 instead of in the building directory. This caused the build to fail if the
3185 source directory was different from the building directory, and was
3186 read-only.
3187
3188 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3189 file. No doubt somebody will tell me if they don't make sense... Also added
3190 Dan Mooney's comments about building on OpenVMS.
3191
3192 19. Added support for partial matching via the PCRE_PARTIAL option for
3193 pcre_exec() and the \P data escape in pcretest.
3194
3195 20. Extended pcretest with 3 new pattern features:
3196
3197 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3198 write the compiled pattern to the file whose name is "rest-of-line".
3199 This is a straight binary dump of the data, with the saved pointer to
3200 the character tables forced to be NULL. The study data, if any, is
3201 written too. After writing, pcretest reads a new pattern.
3202
3203 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3204 compiled pattern from the given file. There must not be any
3205 occurrences of "<" in the file name (pretty unlikely); if there are,
3206 pcretest will instead treat the initial "<" as a pattern delimiter.
3207 After reading in the pattern, pcretest goes on to read data lines as
3208 usual.
3209
3210 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3211 and 16-bit fields in a compiled pattern, to simulate a pattern that
3212 was compiled on a host of opposite endianness.
3213
3214 21. The pcre-exec() function can now cope with patterns that were compiled on
3215 hosts of opposite endianness, with this restriction:
3216
3217 As for any compiled expression that is saved and used later, the tables
3218 pointer field cannot be preserved; the extra_data field in the arguments
3219 to pcre_exec() should be used to pass in a tables address if a value
3220 other than the default internal tables were used at compile time.
3221
3222 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3223 now diagnosed as an error. Previously, most of the time, a negative number
3224 would have been treated as zero, but if in addition "ovector" was passed as
3225 NULL, a crash could occur.
3226
3227 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3228 new versions from the libtool 1.5 distribution (the last one is a copy of
3229 a file called libtool.m4). This seems to have fixed the need to patch
3230 "configure" to support Darwin 1.3 (which I used to do). However, I still
3231 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3232 workstation).
3233
3234 24. Changed the PCRE licence to be the more standard "BSD" licence.
3235
3236
3237 Version 4.5 01-Dec-03
3238 ---------------------
3239
3240 1. There has been some re-arrangement of the code for the match() function so
3241 that it can be compiled in a version that does not call itself recursively.
3242 Instead, it keeps those local variables that need separate instances for
3243 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3244 needs to "recurse". Keeping track of where control must go is done by means
3245 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3246 hide most of the details from the main code, and operates only if
3247 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3248 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3249 operating.
3250
3251 To make it easier for callers to provide specially tailored get/free
3252 functions for this usage, two new functions, pcre_stack_malloc, and
3253 pcre_stack_free, are used. They are always called in strict stacking order,
3254 and the size of block requested is always the same.
3255
3256 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3257 PCRE has been compiled to use the stack or the heap for recursion. The
3258 -C option of pcretest uses this to show which version is compiled.
3259
3260 A new data escape \S, is added to pcretest; it causes the amounts of store
3261 obtained and freed by both kinds of malloc/free at match time to be added
3262 to the output.
3263
3264 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3265 what's available on my current Linux desktop machine.
3266
3267 3. When matching a UTF-8 string, the test for a valid string at the start has
3268 been extended. If start_offset is not zero, PCRE now checks that it points
3269 to a byte that is the start of a UTF-8 character. If not, it returns
3270 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3271 this is necessary because there may be backward assertions in the pattern.
3272 When matching the same subject several times, it may save resources to use
3273 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3274
3275 4. The code for checking the validity of UTF-8 strings has been tightened so
3276 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3277 containing "overlong sequences".
3278
3279 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3280 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3281 should have been "&", but it just so happened that all the cases this let
3282 through by mistake were picked up later in the function.
3283
3284 6. I had used a variable called "isblank" - this is a C99 function, causing
3285 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3286
3287 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3288 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3289 so the version is shown; (c) stop "make test" echoing "./RunTest".
3290
3291 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3292
3293 9. The prototype for memmove() for systems that don't have it was using
3294 size_t, but the inclusion of the header that defines size_t was later. I've
3295 moved the #includes for the C headers earlier to avoid this.
3296
3297 10. Added some adjustments to the code to make it easier to compiler on certain
3298 special systems:
3299
3300 (a) Some "const" qualifiers were missing.
3301 (b) Added the macro EXPORT before all exported functions; by default this
3302 is defined to be empty.
3303 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3304 that it reads its output file name as an argument instead of writing
3305 to the standard output and assuming this can be redirected.
3306
3307 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3308 class containing characters with values greater than 255, PCRE compilation
3309 went into a loop.
3310
3311 12. A recursive reference to a subpattern that was within another subpattern
3312 that had a minimum quantifier of zero caused PCRE to crash. For example,
3313 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3314 recursion. If the recursively-called subpattern itself had a zero repeat,
3315 that was OK.
3316
3317 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3318 buffer into which it was copied (for escape processing) was still set at
3319 1024, so long lines caused crashes.
3320
3321 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3322 "internal error: code overflow...". This applied to any character class
3323 that was followed by a possessive quantifier.
3324
3325 15. Modified the Makefile to add libpcre.la as a prerequisite for
3326 libpcreposix.la because I was told this is needed for a parallel build to
3327 work.
3328
3329 16. If a pattern that contained .* following optional items at the start was
3330 studied, the wrong optimizing data was generated, leading to matching
3331 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3332 matching string must start with a or b or c. The correct conclusion for
3333 this pattern is that a match can start with any character.
3334
3335
3336 Version 4.4 13-Aug-03
3337 ---------------------
3338
3339 1. In UTF-8 mode, a character class containing characters with values between
3340 127 and 255 was not handled correctly if the compiled pattern was studied.
3341 In fixing this, I have also improved the studying algorithm for such
3342 classes (slightly).
3343
3344 2. Three internal functions had redundant arguments passed to them. Removal
3345 might give a very teeny performance improvement.
3346
3347 3. Documentation bug: the value of the capture_top field in a callout is *one
3348 more than* the number of the hightest numbered captured substring.
3349
3350 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3351 in incorrectly linking with a previously installed version. They now link
3352 explicitly with libpcre.la.
3353
3354 5. configure.in no longer needs to recognize Cygwin specially.
3355
3356 6. A problem in pcre.in for Windows platforms is fixed.
3357
3358 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3359 pcretest, it used to include the size of the study block as part of its
3360 output. Unfortunately, the structure contains a field that has a different
3361 size on different hardware architectures. This meant that the tests that
3362 showed this size failed. As the block is currently always of a fixed size,
3363 this information isn't actually particularly useful in pcretest output, so
3364 I have just removed it.
3365
3366 8. Three pre-processor statements accidentally did not start in column 1.
3367 Sadly, there are *still* compilers around that complain, even though
3368 standard C has not required this for well over a decade. Sigh.
3369
3370 9. In pcretest, the code for checking callouts passed small integers in the
3371 callout_data field, which is a void * field. However, some picky compilers
3372 complained about the casts involved for this on 64-bit systems. Now
3373 pcretest passes the address of the small integer instead, which should get
3374 rid of the warnings.
3375
3376 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3377 both compile and run time, and gives an error if an invalid UTF-8 sequence
3378 is found. There is a option for disabling this check in cases where the
3379 string is known to be correct and/or the maximum performance is wanted.
3380
3381 11. In response to a bug report, I changed one line in Makefile.in from
3382
3383 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3384 to
3385 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3386
3387 to look similar to other lines, but I have no way of telling whether this
3388 is the right thing to do, as I do not use Windows. No doubt I'll get told
3389 if it's wrong...
3390
3391
3392 Version 4.3 21-May-03
3393 ---------------------
3394
3395 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3396 Makefile.
3397
3398 2. Some refactoring to improve the quality of the code:
3399
3400 (i) The utf8_table... variables are now declared "const".
3401
3402 (ii) The code for \cx, which used the "case flipping" table to upper case
3403 lower case letters, now just substracts 32. This is ASCII-specific,
3404 but the whole concept of \cx is ASCII-specific, so it seems
3405 reasonable.
3406
3407 (iii) PCRE was using its character types table to recognize decimal and
3408 hexadecimal digits in the pattern. This is silly, because it handles
3409 only 0-9, a-f, and A-F, but the character types table is locale-
3410 specific, which means strange things might happen. A private
3411 table is now used for this - though it costs 256 bytes, a table is
3412 much faster than multiple explicit tests. Of course, the standard
3413 character types table is still used for matching digits in subject
3414 strings against \d.
3415
3416 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3417 ending in _t are). So I've renamed it as ESC_tee.
3418
3419 3. The first argument for regexec() in the POSIX wrapper should have been
3420 defined as "const".
3421
3422 4. Changed pcretest to use malloc() for its buffers so that they can be
3423 Electric Fenced for debugging.
3424
3425 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3426 to read one or more bytes before the start of the subject string. Often this
3427 had no effect on PCRE's behaviour, but in some circumstances it could
3428 provoke a segmentation fault.
3429
3430 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3431 to try to read one or more bytes before the start of the subject string.
3432
3433 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3434 UTF-8 support could misbehave in various ways if the subject string
3435 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3436 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3437 back over UTF-8 characters.)
3438
3439
3440 Version 4.2 14-Apr-03
3441 ---------------------
3442
3443 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3444
3445 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3446 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3447 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3448 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3449 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3450 and BUILD_EXEEXT
3451 Note: automatic setting of the BUILD variables is not yet working
3452 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3453 compile-time but not at link-time
3454 [LINK]: use for linking executables only
3455 make different versions for Windows and non-Windows
3456 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3457 libraries
3458 [LINK_FOR_BUILD]: new variable
3459 [OBJEXT]: use throughout
3460 [EXEEXT]: use throughout
3461 <winshared>: new target
3462 <wininstall>: new target
3463 <dftables.o>: use native compiler
3464 <dftables>: use native linker
3465 <install>: handle Windows platform correctly
3466 <clean>: ditto
3467 <check>: ditto
3468 copy DLL to top builddir before testing
3469
3470 As part of these changes, -no-undefined was removed again. This was reported
3471 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3472 in any case.
3473
3474 3. Some tidies to get rid of compiler warnings:
3475
3476 . In the match_data structure, match_limit was an unsigned long int, whereas
3477 match_call_count was an int. I've made them both unsigned long ints.
3478
3479 . In pcretest the fact that a const uschar * doesn't automatically cast to
3480 a void * provoked a warning.
3481
3482 . Turning on some more compiler warnings threw up some "shadow" variables
3483 and a few more missing casts.
3484
3485 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3486 option, a class that contained a single character with a value between 128
3487 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3488
3489 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3490 option, a class that contained several characters, but with at least one
3491 whose value was between 128 and 255 caused PCRE to crash.
3492
3493
3494 Version 4.1 12-Mar-03
3495 ---------------------
3496
3497 1. Compiling with gcc -pedantic found a couple of places where casts were
3498 needed, and a string in dftables.c that was longer than standard compilers are
3499 required to support.
3500
3501 2. Compiling with Sun's compiler found a few more places where the code could
3502 be tidied up in order to avoid warnings.
3503
3504 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3505 first of these names is deprecated in the latest Autoconf in favour of the name
3506 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3507 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3508 analogy I have changed it to CFLAGS_FOR_BUILD.
3509
3510 4. Added -no-undefined to the linking command in the Makefile, because this is
3511 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3512 linking step for the pcreposix library.
3513
3514 5. PCRE was failing to diagnose the case of two named groups with the same
3515 name.
3516
3517 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3518 literal character that is needed in the subject for a match, and scans along to
3519 ensure that it is present before embarking on the full matching process. This
3520 saves time in cases of nested unlimited repeats that are never going to match.
3521 Problem: the scan can take a lot of time if the subject is very long (e.g.
3522 megabytes), thus penalizing straightforward matches. It is now done only if the
3523 amount of subject to be scanned is less than 1000 bytes.
3524
3525 7. A lesser problem with the same optimization is that it was recording the
3526 first character of an anchored pattern as "needed", thus provoking a search
3527 right along the subject, even when the first match of the pattern was going to
3528 fail. The "needed" character is now not set for anchored patterns, unless it
3529 follows something in the pattern that is of non-fixed length. Thus, it still
3530 fulfils its original purpose of finding quick non-matches in cases of nested
3531 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3532
3533
3534 Version 4.0 17-Feb-03
3535 ---------------------
3536
3537 1. If a comment in an extended regex that started immediately after a meta-item
3538 extended to the end of string, PCRE compiled incorrect data. This could lead to
3539 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3540
3541 2. Moved to autoconf 2.53 and libtool 1.4.2.
3542
3543 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3544 the special perltest8 script is no longer needed - all the tests can be run
3545 from a single perltest script.
3546
3547 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3548 by \s. It has now been removed in PCRE. This means it isn't recognized as
3549 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3550 class [:space:] *does* include VT, thereby creating a mess.
3551
3552 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3553 space and tab.
3554
3555 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3556 its new features into the main test script, reducing the number of scripts.
3557
3558 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3559 were backward compatible, and made the (?i) apply to the whole pattern, as if
3560 /i were given. Now it behaves more logically, and applies the option setting
3561 only to what follows. PCRE has been changed to follow suit. However, if it
3562 finds options settings right at the start of the pattern, it extracts them into
3563 the global options, as before. Thus, they show up in the info data.
3564
3565 8. Added support for the \Q...\E escape sequence. Characters in between are
3566 treated as literals. This is slightly different from Perl in that $ and @ are
3567 also handled as literals inside the quotes. In Perl, they will cause variable
3568 interpolation. Note the following examples:
3569
3570 Pattern PCRE matches Perl matches
3571
3572 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3573 \Qabc\$xyz\E abc\$xyz abc\$xyz
3574 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3575
3576 For compatibility with Perl, \Q...\E sequences are recognized inside character
3577 classes as well as outside them.
3578
3579 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3580 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3581 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3582 signed/unsigned warnings.
3583
3584 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3585 option for pcretest, so I've replaced it by a simple function that does just
3586 that job.
3587
3588 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3589 "pcregrep -".
3590
3591 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3592 Java package. This provides some syntactic sugar for simple cases of what my
3593 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3594 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3595 item, you can use this simplified notation. Note that only makes sense with
3596 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3597 greediness, whatever the setting of the PCRE_UNGREEDY option.
3598
3599 13. A change of greediness default within a pattern was not taking effect at
3600 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3601 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3602 was abstracted outside.
3603
3604 14. PCRE now supports the \G assertion. It is true when the current matching
3605 position is at the start point of the match. This differs from \A when the
3606 starting offset is non-zero. Used with the /g option of pcretest (or similar
3607 code), it works in the same way as it does for Perl's /g option. If all
3608 alternatives of a regex begin with \G, the expression is anchored to the start
3609 match position, and the "anchored" flag is set in the compiled expression.
3610
3611 15. Some bugs concerning the handling of certain option changes within patterns
3612 have been fixed. These applied to options other than (?ims). For example,
3613 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3614 been the other way round. Some of this was related to change 7 above.
3615
3616 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3617 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3618 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3619 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3620
3621 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3622 mode. Unlike ".", it always matches newline, whatever the setting of
3623 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3624 assertions. Perl allows it, but it doesn't (in general) work because it can't
3625 calculate the length of the lookbehind. At least, that's the case for Perl
3626 5.8.0 - I've been told they are going to document that it doesn't work in
3627 future.
3628
3629 18. Added an error diagnosis for escapes that PCRE does not support: these are
3630 \L, \l, \N, \P, \p, \U, \u, and \X.
3631
3632 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3633 reading past the end of the pattern in cases such as /[abcd/.
3634
3635 20. PCRE was getting more memory than necessary for patterns with classes that
3636 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3637
3638 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3639 compiling PCRE for use with Virtual Pascal.
3640
3641 22. Small fix to the Makefile to make it work properly if the build is done
3642 outside the source tree.
3643
3644 23. Added a new extension: a condition to go with recursion. If a conditional
3645 subpattern starts with (?(R) the "true" branch is used if recursion has
3646 happened, whereas the "false" branch is used only at the top level.
3647
3648 24. When there was a very long string of literal characters (over 255 bytes
3649 without UTF support, over 250 bytes with UTF support), the computation of how
3650 much memory was required could be incorrect, leading to segfaults or other
3651 strange effects.
3652
3653 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3654 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3655 there was a subsequent back reference to those brackets. This meant that, for
3656 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3657 possible to check for precisely this case. All we can do is abandon the
3658 optimization if .* occurs inside capturing brackets when there are any back
3659 references whatsoever. (See below for a better fix that came later.)
3660
3661 26. The handling of the optimization for finding the first character of a
3662 non-anchored pattern, and for finding a character that is required later in the
3663 match were failing in some cases. This didn't break the matching; it just
3664 failed to optimize when it could. The way this is done has been re-implemented.
3665
3666 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3667
3668 28. Added a new feature that provides some of the functionality that Perl
3669 provides with (?{...}). The facility is termed a "callout". The way it is done
3670 in PCRE is for the caller to provide an optional function, by setting
3671 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3672 global variable. By default it is unset, which disables all calling out. To get
3673 the function called, the regex must include (?C) at appropriate points. This
3674 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3675 This provides a means of identifying different callout points. When PCRE
3676 reaches such a point in the regex, if pcre_callout has been set, the external
3677 function is called. It is provided with data in a structure called
3678 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3679 matching continues; if it returns a non-zero value, the match at the current
3680 point fails. However, backtracking will occur if possible. [This was changed
3681 later and other features added - see item 49 below.]
3682
3683 29. pcretest is upgraded to test the callout functionality. It provides a
3684 callout function that displays information. By default, it shows the start of
3685 the match and the current position in the text. There are some new data escapes
3686 to vary what happens:
3687
3688 \C+ in addition, show current contents of captured substrings
3689 \C- do not supply a callout function
3690 \C!n return 1 when callout number n is reached
3691 \C!n!m return 1 when callout number n is reached for the mth time
3692
3693 30. If pcregrep was called with the -l option and just a single file name, it
3694 output "<stdin>" if a match was found, instead of the file name.
3695
3696 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3697 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3698 pcre_exec(). This saves a malloc/free per call. The default value of
3699 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3700 when configuring.
3701
3702 32. The default maximum size of a compiled pattern is 64K. There have been a
3703 few cases of people hitting this limit. The code now uses macros to handle the
3704 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3705 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3706 configuring. Tests 2 and 5 work only with 2-byte links because they output
3707 debugging information about compiled patterns.
3708
3709 33. Internal code re-arrangements:
3710
3711 (a) Moved the debugging function for printing out a compiled regex into
3712 its own source file (printint.c) and used #include to pull it into
3713 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3714 separate copies.
3715
3716 (b) Defined the list of op-code names for debugging as a macro in
3717 internal.h so that it is next to the definition of the opcodes.
3718
3719 (c) Defined a table of op-code lengths for simpler skipping along compiled
3720 code. This is again a macro in internal.h so that it is next to the
3721 definition of the opcodes.
3722
3723 34. Added support for recursive calls to individual subpatterns, along the
3724 lines of Robin Houston's patch (but implemented somewhat differently).
3725
3726 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3727 allow it to read and process whole directories in Win32. This code was
3728 contributed by Lionel Fourquaux; it has not been tested by me.
3729
3730 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3731 used to name a group. Names consist of alphanumerics and underscores, and must
3732 be unique. Back references use the syntax (?P=name) and recursive calls use
3733 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3734 numbers. The function pcre_fullinfo() can be used after compilation to extract
3735 a name/number map. There are three relevant calls:
3736
3737 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3738 PCRE_INFO_NAMECOUNT yields the number of entries
3739 PCRE_INFO_NAMETABLE yields a pointer to the map.
3740
3741 The map is a vector of fixed-size entries. The size of each entry depends on
3742 the length of the longest name used. The first two bytes of each entry are the
3743 group number, most significant byte first. There follows the corresponding
3744 name, zero terminated. The names are in alphabetical order.
3745
3746 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3747 case instead of 255. Making it the same both with and without UTF-8 support
3748 means that the same test output works with both.
3749
3750 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3751 calling malloc() with a zero argument.
3752
3753 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3754 optimization. I've improved things by keeping a bitmap of backreferences with
3755 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3756 fact referenced, the optimization can be applied. It is unlikely that a
3757 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3758 the match to follow \n) will appear inside brackets with a number greater than
3759 31, but if it does, any back reference > 31 suppresses the optimization.
3760
3761 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3762 of disabling numbered capturing parentheses. Any opening parenthesis that is
3763 not followed by ? behaves as if it were followed by ?: but named parentheses
3764 can still be used for capturing (and they will acquire numbers in the usual
3765 way).
3766
3767 41. Redesigned the return codes from the match() function into yes/no/error so
3768 that errors can be passed back from deep inside the nested calls. A malloc
3769 failure while inside a recursive subpattern call now causes the
3770 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3771
3772 42. It is now possible to set a limit on the number of times the match()
3773 function is called in a call to pcre_exec(). This facility makes it possible to
3774 limit the amount of recursion and backtracking, though not in a directly
3775 obvious way, because the match() function is used in a number of different
3776 circumstances. The count starts from zero for each position in the subject
3777 string (for non-anchored patterns). The default limit is, for compatibility, a
3778 large number, namely 10 000 000. You can change this in two ways:
3779
3780 (a) When configuring PCRE before making, you can use --with-match-limit=n
3781 to set a default value for the compiled library.
3782
3783 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3784 a different value is set. See 45 below.
3785
3786 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3787
3788 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3789 of things that can be changed at compile time. The first argument specifies
3790 what is wanted and the second points to where the information is to be placed.
3791 The current list of available information is:
3792
3793 PCRE_CONFIG_UTF8
3794
3795 The output is an integer that is set to one if UTF-8 support is available;
3796 otherwise it is set to zero.
3797
3798 PCRE_CONFIG_NEWLINE
3799
3800 The output is an integer that it set to the value of the code that is used for
3801 newline. It is either LF (10) or CR (13).
3802
3803 PCRE_CONFIG_LINK_SIZE
3804
3805 The output is an integer that contains the number of bytes used for internal
3806 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3807
3808 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3809
3810 The output is an integer that contains the threshold above which the POSIX
3811 interface uses malloc() for output vectors. See item 31 above.
3812
3813 PCRE_CONFIG_MATCH_LIMIT
3814
3815 The output is an unsigned integer that contains the default limit of the number
3816 of match() calls in a pcre_exec() execution. See 42 above.
3817
3818 44. pcretest has been upgraded by the addition of the -C option. This causes it
3819 to extract all the available output from the new pcre_config() function, and to
3820 output it. The program then exits immediately.
3821
3822 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3823 order to support additional features. One way would have been to define
3824 pcre_exec2() (for example) with extra arguments, but this would not have been
3825 extensible, and would also have required all calls to the original function to
3826 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3827 is used for passing in "extra" data from pcre_study().
3828
3829 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3830 contains the following fields:
3831
3832 flags a bitmap indicating which of the following fields are set
3833 study_data opaque data from pcre_study()
3834 match_limit a way of specifying a limit on match() calls for a specific
3835 call to pcre_exec()
3836 callout_data data for callouts (see 49 below)
3837
3838 The flag bits are also defined in pcre.h, and are
3839
3840 PCRE_EXTRA_STUDY_DATA
3841 PCRE_EXTRA_MATCH_LIMIT
3842 PCRE_EXTRA_CALLOUT_DATA
3843
3844 The pcre_study() function now returns one of these new pcre_extra blocks, with
3845 the actual study data pointed to by the study_data field, and the
3846 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3847 before. That is, this change is entirely upwards-compatible and requires no
3848 change to existing code.
3849
3850 If you want to pass in additional data to pcre_exec(), you can either place it
3851 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3852 block.
3853
3854 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3855 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3856 times with different match limits, until it finds the minimum value needed for
3857 pcre_exec() to complete. The value is then output. This can be instructive; for
3858 most simple matches the number is quite small, but for pathological cases it
3859 gets very large very quickly.
3860
3861 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3862 returns the size of the data block pointed to by the study_data field in a
3863 pcre_extra block, that is, the value that was passed as the argument to
3864 pcre_malloc() when PCRE was getting memory in which to place the information
3865 created by pcre_study(). The fourth argument should point to a size_t variable.
3866 pcretest has been extended so that this information is shown after a successful
3867 pcre_study() call when information about the compiled regex is being displayed.
3868
3869 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3870 because what follows is always an absolute path. (Later: it turns out that this
3871 is more than cosmetic for MinGW, because it doesn't like empty path
3872 components.)
3873
3874 49. Some changes have been made to the callout feature (see 28 above):
3875
3876 (i) A callout function now has three choices for what it returns:
3877
3878 0 => success, carry on matching
3879 > 0 => failure at this point, but backtrack if possible
3880 < 0 => serious error, return this value from pcre_exec()
3881
3882 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3883 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3884 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3885 use by callout functions. It will never be used by PCRE itself.
3886
3887 (ii) The pcre_extra structure (see 45 above) has a void * field called
3888 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3889 pcre_callout_block structure has a field of the same name. The contents of
3890 the field passed in the pcre_extra structure are passed to the callout
3891 function in the corresponding field in the callout block. This makes it
3892 easier to use the same callout-containing regex from multiple threads. For
3893 testing, the pcretest program has a new data escape
3894
3895 \C*n pass the number n (may be negative) as callout_data
3896
3897 If the callout function in pcretest receives a non-zero value as
3898 callout_data, it returns that value.
3899
3900 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3901 there were some redundant $(CFLAGS) in commands that are now specified as
3902 $(LINK), which already includes $(CFLAGS).
3903
3904 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3905 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3906 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3907 one-byte characters throughout. Note that case-insensitive matching applies
3908 only to characters whose values are less than 256. PCRE doesn't support the
3909 notion of cases for higher-valued characters.
3910
3911 (i) A character class whose characters are all within 0-255 is handled as
3912 a bit map, and the map is inverted for negative classes. Previously, a
3913 character > 255 always failed to match such a class; however it should
3914 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3915
3916 (ii) A negated character class with a single character < 255 is coded as
3917 "not this character" (OP_NOT). This wasn't working properly when the test
3918 character was multibyte, either singly or repeated.
3919
3920 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3921 mode, for example: \x{100}{2,3}.
3922
3923 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3924 singly or repeated) now correctly test multibyte characters. However,
3925 PCRE doesn't recognize any characters with values greater than 255 as
3926 digits, spaces, or word characters. Such characters always match \D, \S,
3927 and \W, and never match \d, \s, or \w.
3928
3929 (v) Classes may now contain characters and character ranges with values
3930 greater than 255. For example: [ab\x{100}-\x{400}].
3931
3932 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3933 PCRE in UTF-8 mode.
3934
3935 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3936 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3937 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3938 value.)
3939
3940 53. The single man page has become too large. I have therefore split it up into
3941 a number of separate man pages. These also give rise to individual HTML pages;
3942 these are now put in a separate directory, and there is an index.html page that
3943 lists them all. Some hyperlinking between the pages has been installed.
3944
3945 54. Added convenience functions for handling named capturing parentheses.
3946
3947 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3948 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3949 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3950 are faulted.
3951
3952 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3953 calling configure. These values are used when compiling the dftables.c program
3954 which is run to generate the source of the default character tables. They
3955 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3956 you will need to set these values.
3957
3958 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3959
3960
3961 Version 3.9 02-Jan-02
3962 ---------------------
3963
3964 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3965
3966 2. If --disable-static was given, the building process failed when trying to
3967 build pcretest and pcregrep. (For some reason it was using libtool to compile
3968 them, which is not right, as they aren't part of the library.)
3969
3970
3971 Version 3.8 18-Dec-01
3972 ---------------------
3973
3974 1. The experimental UTF-8 code was completely screwed up. It was packing the
3975 bytes in the wrong order. How dumb can you get?
3976
3977
3978 Version 3.7 29-Oct-01
3979 ---------------------
3980
3981 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3982 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3983 this didn't happen under Solaris 8, where I normally test things.
3984
3985 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3986 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3987 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3988 this fix an hour or so after the initial 3.7 release.)
3989
3990
3991 Version 3.6 23-Oct-01
3992 ---------------------
3993
3994 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3995 offsets passed as NULL with zero offset count.
3996
3997 2. The config.guess and config.sub files had not been updated when I moved to
3998 the latest autoconf.
3999
4000
4001 Version 3.5 15-Aug-01
4002 ---------------------
4003
4004 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4005 had been forgotten.
4006
4007 2. By using declared but undefined structures, we can avoid using "void"
4008 definitions in pcre.h while keeping the internal definitions of the structures
4009 private.
4010
4011 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4012 user point of view, this means that both static and shared libraries are built
4013 by default, but this can be individually controlled. More of the work of
4014 handling this static/shared cases is now inside libtool instead of PCRE's make
4015 file.
4016
4017 4. The pcretest utility is now installed along with pcregrep because it is
4018 useful for users (to test regexs) and by doing this, it automatically gets
4019 relinked by libtool. The documentation has been turned into a man page, so
4020 there are now .1, .txt, and .html versions in /doc.
4021
4022 5. Upgrades to pcregrep:
4023 (i) Added long-form option names like gnu grep.
4024 (ii) Added --help to list all options with an explanatory phrase.
4025 (iii) Added -r, --recursive to recurse into sub-directories.
4026 (iv) Added -f, --file to read patterns from a file.
4027
4028 6. pcre_exec() was referring to its "code" argument before testing that
4029 argument for NULL (and giving an error if it was NULL).
4030
4031 7. Upgraded Makefile.in to allow for compiling in a different directory from
4032 the source directory.
4033
4034 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4035 options bits, the pointer it was passed was to an int instead of to an unsigned
4036 long int. This mattered only on 64-bit systems.
4037
4038 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4039 generated) instead of pcre.in, which it its source. Also made the same change
4040 in several of the .c files.
4041
4042 10. A new release of gcc defines printf() as a macro, which broke pcretest
4043 because it had an ifdef in the middle of a string argument for printf(). Fixed
4044 by using separate calls to printf().
4045
4046 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4047 script, to force use of CR or LF instead of \n in the source. On non-Unix
4048 systems, the value can be set in config.h.
4049
4050 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4051 absolute limit. Changed the text of the error message to make this clear, and
4052 likewise updated the man page.
4053
4054 13. The limit of 99 on the number of capturing subpatterns has been removed.
4055 The new limit is 65535, which I hope will not be a "real" limit.
4056
4057
4058 Version 3.4 22-Aug-00
4059 ---------------------
4060
4061 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4062
4063 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4064
4065
4066 Version 3.3 01-Aug-00
4067 ---------------------
4068
4069 1. If an octal character was given, but the value was greater than \377, it
4070 was not getting masked to the least significant bits, as documented. This could
4071 lead to crashes in some systems.
4072
4073 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4074 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4075
4076 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4077 These just pass their arguments on to (pcre_free)(), but they are provided
4078 because some uses of PCRE bind it to non-C systems that can call its functions,
4079 but cannot call free() or pcre_free() directly.
4080
4081 4. Add "make test" as a synonym for "make check". Corrected some comments in
4082 the Makefile.
4083
4084 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4085 Makefile.
4086
4087 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4088 command called pgrep for grepping around the active processes.
4089
4090 7. Added the beginnings of support for UTF-8 character strings.
4091
4092 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4093 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4094 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4095 out for the ar command.)
4096
4097
4098 Version 3.2 12-May-00
4099 ---------------------
4100
4101 This is purely a bug fixing release.
4102
4103 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4104 of ZA. This was just one example of several cases that could provoke this bug,
4105 which was introduced by change 9 of version 2.00. The code for breaking
4106 infinite loops after an iteration that matches an empty string was't working
4107 correctly.
4108
4109 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4110 when matched against abbab (for example). After matching an empty string, it
4111 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4112 caused it to match further down the string than it should.
4113
4114 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4115 was there because it doesn't seem to be needed, and it causes trouble on some
4116 systems, as it is not a Standard C header. It has been removed.
4117
4118 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4119 were reported on the Macintosh. The changes were from
4120
4121 while ((c = *(++ptr)) != 0 && c != '\n');
4122 to
4123 while ((c = *(++ptr)) != 0 && c != '\n') ;
4124
4125 Totally extraordinary, but if that's what it takes...
4126
4127 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4128 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4129 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4130 assumes the way PCRE uses memmove() (always moving upwards).
4131
4132 6. PCRE is being used in one environment where strchr() is not available. There
4133 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4134 faster code anyway.
4135
4136
4137 Version 3.1 09-Feb-00
4138 ---------------------
4139
4140 The only change in this release is the fixing of some bugs in Makefile.in for
4141 the "install" target:
4142
4143 (1) It was failing to install pcreposix.h.
4144
4145 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4146
4147
4148 Version 3.0 01-Feb-00
4149 ---------------------
4150
4151 1. Add support for the /+ modifier to perltest (to output $` like it does in
4152 pcretest).
4153
4154 2. Add support for the /g modifier to perltest.
4155
4156 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4157 matches null strings.
4158
4159 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4160 pattern. Perl treats empty patterns specially - it reuses the most recent
4161 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4162 effect.
4163
4164 5. The POSIX interface was broken in that it was just handing over the POSIX
4165 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4166 required a bigger vector, with some working space on the end. This means that
4167 the POSIX wrapper now has to get and free some memory, and copy the results.
4168
4169 6. Added some simple autoconf support, placing the test data and the
4170 documentation in separate directories, re-organizing some of the
4171 information files, and making it build pcre-config (a GNU standard). Also added
4172 libtool support for building PCRE as a shared library, which is now the
4173 default.
4174
4175 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4176 09 are not valid octal constants. Single digits will be used for minor values
4177 less than 10.
4178
4179 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4180 existing programs that set these in the POSIX interface can use PCRE without
4181 modification.
4182
4183 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4184 return all that pcre_info() returns, plus additional data. The pcre_info()
4185 function is retained for compatibility, but is considered to be obsolete.
4186
4187 10. Added experimental recursion feature (?R) to handle one common case that
4188 Perl 5.6 will be able to do with (?p{...}).
4189
4190 11. Added support for POSIX character classes like [:alpha:], which Perl is
4191 adopting.
4192
4193
4194 Version 2.08 31-Aug-99
4195 ----------------------
4196
4197 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4198 trying to match at the startoffset position, but instead was moving forward to
4199 the next newline as if a previous match had failed.
4200
4201 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4202 and could get into a loop if a null string was matched other than at the start
4203 of the subject.
4204
4205 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4206 be distinguished at compile time, and for completeness also added PCRE_DATE.
4207
4208 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4209 in GnuWin32 environments.
4210
4211
4212 Version 2.07 29-Jul-99
4213 ----------------------
4214
4215 1. The documentation is now supplied in plain text form and HTML as well as in
4216 the form of man page sources.
4217
4218 2. C++ compilers don't like assigning (void *) values to other pointer types.
4219 In particular this affects malloc(). Although there is no problem in Standard
4220 C, I've put in casts to keep C++ compilers happy.
4221
4222 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4223 should be (const char *).
4224
4225 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4226 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4227 However, I haven't made this a standard facility. The documentation doesn't
4228 mention it, and the Makefile doesn't support it.
4229
4230 5. The Makefile now contains an "install" target, with editable destinations at
4231 the top of the file. The pcretest program is not installed.
4232
4233 6. pgrep -V now gives the PCRE version number and date.
4234
4235 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4236 causing the entire string to be ignored, instead of just the last character.
4237
4238 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4239 non-matching string, it can take a very, very long time, even for strings of
4240 quite modest length, because of the nested recursion. PCRE now does better in
4241 some of these cases. It does this by remembering the last required literal
4242 character in the pattern, and pre-searching the subject to ensure it is present
4243 before running the real match. In other words, it applies a heuristic to detect
4244 some types of certain failure quickly, and in the above example, if presented
4245 with a string that has no trailing " it gives "no match" very quickly.
4246
4247 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4248 other alternatives are tried instead.
4249
4250
4251 Version 2.06 09-Jun-99
4252 ----------------------
4253
4254 1. Change pcretest's output for amount of store used to show just the code
4255 space, because the remainder (the data block) varies in size between 32-bit and
4256 64-bit systems.
4257
4258 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4259 start matching at. This allows lookbehinds to work when searching for multiple
4260 occurrences in a string.
4261
4262 3. Added additional options to pcretest for testing multiple occurrences:
4263
4264 /+ outputs the rest of the string that follows a match
4265 /g loops for multiple occurrences, using the new startoffset argument
4266 /G loops for multiple occurrences by passing an incremented pointer
4267
4268 4. PCRE wasn't doing the "first character" optimization for patterns starting
4269 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4270 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4271 the letter 'x'. On long subject strings, this gives a significant speed-up.
4272
4273
4274 Version 2.05 21-Apr-99
4275 ----------------------
4276
4277 1. Changed the type of magic_number from int to long int so that it works
4278 properly on 16-bit systems.
4279
4280 2. Fixed a bug which caused patterns starting with .* not to work correctly
4281 when the subject string contained newline characters. PCRE was assuming
4282 anchoring for such patterns in all cases, which is not correct because .* will
4283 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4284 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4285 must be retried after every newline in the subject.
4286
4287
4288 Version 2.04 18-Feb-99
4289 ----------------------
4290
4291 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4292 computation of the store needed to hold the pattern was incorrect (too large).
4293 If such patterns were nested a few deep, this could multiply and become a real
4294 problem.
4295
4296 2. Added /M option to pcretest to show the memory requirement of a specific
4297 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4298
4299 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4300 compiled in such a way that the backtracking after subsequent failure was
4301 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4302 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4303
4304
4305 Version 2.03 02-Feb-99
4306 ----------------------
4307
4308 1. Fixed typo and small mistake in man page.
4309
4310 2. Added 4th condition (GPL supersedes if conflict) and created separate
4311 LICENCE file containing the conditions.
4312
4313 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4314 Perl, that is the internal \ allows the delimiter to be included in the
4315 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4316 the final delimiter, add \ to the end of the pattern (to test the error).
4317
4318 4. Added the convenience functions for extracting substrings after a successful
4319 match. Updated pcretest to make it able to test these functions.
4320
4321
4322 Version 2.02 14-Jan-99
4323 ----------------------
4324
4325 1. Initialized the working variables associated with each extraction so that
4326 their saving and restoring doesn't refer to uninitialized store.
4327
4328 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4329 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4330 fix the problem.
4331
4332 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4333 calls, and wasn't printing the correct value for compiling calls. Increased the
4334 default value of LOOPREPEAT, and the number of significant figures in the
4335 times.
4336
4337 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4338
4339 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4340 a building problem on Windows NT with a FAT file system.
4341
4342
4343 Version 2.01 21-Oct-98
4344 ----------------------
4345
4346 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4347 to character tables built by pcre_maketables() in the current locale. If NULL
4348 is passed, the default tables are used.
4349
4350
4351 Version 2.00 24-Sep-98
4352 ----------------------
4353
4354 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4355 it any more.
4356
4357 2. Allow quantification of (?>) groups, and make it work correctly.
4358
4359 3. The first character computation wasn't working for (?>) groups.
4360
4361 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4362 end of the subject) and add 5.005's \z, which really does match only at the
4363 very end of the subject.
4364
4365 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4366
4367 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4368 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4369 localized options. All options to pcre_study() were also removed.
4370
4371 7. Add other new features from 5.005:
4372
4373 $(?<= positive lookbehind
4374 $(?<! negative lookbehind
4375 (?imsx-imsx) added the unsetting capability
4376 such a setting is global if at outer level; local otherwise
4377 (?imsx-imsx:) non-capturing groups with option setting
4378 (?(cond)re|re) conditional pattern matching
4379
4380 A backreference to itself in a repeated group matches the previous
4381 captured string.
4382
4383 8. General tidying up of studying (both automatic and via "study")
4384 consequential on the addition of new assertions.
4385
4386 9. As in 5.005, unlimited repeated groups that could match an empty substring
4387 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4388 runtime if any iteration does actually match an empty substring.
4389
4390 10. Include the RunTest script in the distribution.
4391
4392 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4393 discrepancies, some of which were old and were also with respect to 5.004. They
4394 have now been fixed.
4395
4396
4397 Version 1.09 28-Apr-98
4398 ----------------------
4399
4400 1. A negated single character class followed by a quantifier with a minimum
4401 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4402 program crashes, or just wrong answers. This did not apply to negated classes
4403 containing more than one character, or to minima other than one.
4404
4405
4406 Version 1.08 27-Mar-98
4407 ----------------------
4408
4409 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4410
4411 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4412 latter must appear before anything that relies on it in the pattern.
4413
4414
4415 Version 1.07 16-Feb-98
4416 ----------------------
4417
4418 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4419 repeat of a potentially empty string).
4420
4421
4422 Version 1.06 23-Jan-98
4423 ----------------------
4424
4425 1. Added Markus Oberhumer's little patches for C++.
4426
4427 2. Literal strings longer than 255 characters were broken.
4428
4429
4430 Version 1.05 23-Dec-97
4431 ----------------------
4432
4433 1. Negated character classes containing more than one character were failing if
4434 PCRE_CASELESS was set at run time.
4435
4436
4437 Version 1.04 19-Dec-97
4438 ----------------------
4439
4440 1. Corrected the man page, where some "const" qualifiers had been omitted.
4441
4442 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4443 input syntax.
4444
4445 3. Fixed memory leak which occurred when a regex with back references was
4446 matched with an offsets vector that wasn't big enough. The temporary memory
4447 that is used in this case wasn't being freed if the match failed.
4448
4449 4. Tidied pcretest to ensure it frees memory that it gets.
4450
4451 5. Temporary memory was being obtained in the case where the passed offsets
4452 vector was exactly big enough.
4453
4454 6. Corrected definition of offsetof() from change 5 below.
4455
4456 7. I had screwed up change 6 below and broken the rules for the use of
4457 setjmp(). Now fixed.
4458
4459
4460 Version 1.03 18-Dec-97
4461 ----------------------
4462
4463 1. A erroneous regex with a missing opening parenthesis was correctly
4464 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4465 on some systems.
4466
4467 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4468 it was reported that one broken compiler failed on the former because "code" is
4469 also an independent variable.
4470
4471 3. The erroneous regex a[]b caused an array overrun reference.
4472
4473 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4474 fail on data ending with that character. (It was going on too far, and checking
4475 the next character, typically a binary zero.) This was specific to the
4476 optimized code for single-character negative classes.
4477
4478 5. Added a contributed patch from the TIN world which does the following:
4479
4480 + Add an undef for memmove, in case the the system defines a macro for it.
4481
4482 + Add a definition of offsetof(), in case there isn't one. (I don't know
4483 the reason behind this - offsetof() is part of the ANSI standard - but
4484 it does no harm).
4485
4486 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4487 most of the places where whitespace preceded '#'. I have given up and
4488 allowed the remaining 2 cases to be at the margin.
4489
4490 + Rename some variables in pcre to eliminate shadowing. This seems very
4491 pedantic, but does no harm, of course.
4492
4493 6. Moved the call to setjmp() into its own function, to get rid of warnings
4494 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4495
4496 7. Constructs such as \d{8,} were compiling into the equivalent of
4497 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4498 outcome, but in this particular case used more store than had been allocated,
4499 which caused the bug to be discovered because it threw up an internal error.
4500
4501 8. The debugging code in both pcre and pcretest for outputting the compiled
4502 form of a regex was going wrong in the case of back references followed by
4503 curly-bracketed repeats.
4504
4505
4506 Version 1.02 12-Dec-97
4507 ----------------------
4508
4509 1. Typos in pcre.3 and comments in the source fixed.
4510
4511 2. Applied a contributed patch to get rid of places where it used to remove
4512 'const' from variables, and fixed some signed/unsigned and uninitialized
4513 variable warnings.
4514
4515 3. Added the "runtest" target to Makefile.
4516
4517 4. Set default compiler flag to -O2 rather than just -O.
4518
4519
4520 Version 1.01 19-Nov-97
4521 ----------------------
4522
4523 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4524 like /([ab]*)*/, that is, for classes with more than one character in them.
4525
4526 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4527 as /((?>a*))*/ (a PCRE_EXTRA facility).
4528
4529
4530 Version 1.00 18-Nov-97
4531 ----------------------
4532
4533 1. Added compile-time macros to support systems such as SunOS4 which don't have
4534 memmove() or strerror() but have other things that can be used instead.
4535
4536 2. Arranged that "make clean" removes the executables.
4537
4538
4539 Version 0.99 27-Oct-97
4540 ----------------------
4541
4542 1. Fixed bug in code for optimizing classes with only one character. It was
4543 initializing a 32-byte map regardless, which could cause it to run off the end
4544 of the memory it had got.
4545
4546 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4547
4548
4549 Version 0.98 22-Oct-97
4550 ----------------------
4551
4552 1. Fixed bug in code for handling temporary memory usage when there are more
4553 back references than supplied space in the ovector. This could cause segfaults.
4554
4555
4556 Version 0.97 21-Oct-97
4557 ----------------------
4558
4559 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4560
4561 2. Optimized negated single characters not to use a bit map.
4562
4563 3. Brought error texts together as macro definitions; clarified some of them;
4564 fixed one that was wrong - it said "range out of order" when it meant "invalid
4565 escape sequence".
4566
4567 4. Changed some char * arguments to const char *.
4568
4569 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4570
4571 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4572 pcretest.
4573
4574
4575 Version 0.96 16-Oct-97
4576 ----------------------
4577
4578 1. Added a simple "pgrep" utility to the distribution.
4579
4580 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4581 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4582 where "ddd" means "one or more decimal digits".
4583
4584 3. Fixed serious bug. If a pattern had a back reference, but the call to
4585 pcre_exec() didn't supply a large enough ovector to record the related
4586 identifying subpattern, the match always failed. PCRE now remembers the number
4587 of the largest back reference, and gets some temporary memory in which to save
4588 the offsets during matching if necessary, in order to ensure that
4589 backreferences always work.
4590
4591 4. Increased the compatibility with Perl in a number of ways:
4592
4593 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4594 to request this handling. The option can be set at compile or exec time.
4595
4596 (b) $ matches before a terminating newline by default; an option
4597 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4598 mode). The option can be set at compile or exec time.
4599
4600 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4601 the same as Perl's. If the decimal number it represents is less than 10
4602 or there aren't that many previous left capturing parentheses, an octal
4603 escape is read. Inside a character class, it's always an octal escape,
4604 even if it is a single digit.
4605
4606 (d) An escaped but undefined alphabetic character is taken as a literal,
4607 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4608 escapes.
4609
4610 (e) {0} is now permitted. (The previous item is removed from the compiled
4611 pattern).
4612
4613 5. Changed all the names of code files so that the basic parts are no longer
4614 than 10 characters, and abolished the teeny "globals.c" file.
4615
4616 6. Changed the handling of character classes; they are now done with a 32-byte
4617 bit map always.
4618
4619 7. Added the -d and /D options to pcretest to make it possible to look at the
4620 internals of compilation without having to recompile pcre.
4621
4622
4623 Version 0.95 23-Sep-97
4624 ----------------------
4625
4626 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4627 \x20 at the start of a run of normal characters. These were being treated as
4628 real characters, instead of the source characters being re-checked.
4629
4630
4631 Version 0.94 18-Sep-97
4632 ----------------------
4633
4634 1. The functions are now thread-safe, with the caveat that the global variables
4635 containing pointers to malloc() and free() or alternative functions are the
4636 same for all threads.
4637
4638 2. Get pcre_study() to generate a bitmap of initial characters for non-
4639 anchored patterns when this is possible, and use it if passed to pcre_exec().
4640
4641
4642 Version 0.93 15-Sep-97
4643 ----------------------
4644
4645 1. /(b)|(:+)/ was computing an incorrect first character.
4646
4647 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4648 but not actually doing anything yet.
4649
4650 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4651 as Perl does (e.g. [-az] or [az-]).
4652
4653 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4654 all possible positions.
4655
4656 5. Split up into different modules to avoid including unneeded functions in a
4657 compiled binary. However, compile and exec are still in one module. The "study"
4658 function is split off.
4659
4660 6. The character tables are now in a separate module whose source is generated
4661 by an auxiliary program - but can then be edited by hand if required. There are
4662 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4663 toupper() in the code.
4664
4665 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4666 make them global. Abolish the function for setting them, as the caller can now
4667 set them directly.
4668
4669
4670 Version 0.92 11-Sep-97
4671 ----------------------
4672
4673 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4674 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4675
4676 2. Caseless matching was not working in character classes if the characters in
4677 the pattern were in upper case.
4678
4679 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4680
4681 4. Make PCRE_ANCHORED public and accept as a compile option.
4682
4683 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4684 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4685 pass them.
4686
4687 6. Give an error if bad option bits passed at compile or run time.
4688
4689 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4690 pcretest to cause it to pass that flag.
4691
4692 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4693 options, and the first character, if set.
4694
4695 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4696
4697
4698 Version 0.91 10-Sep-97
4699 ----------------------
4700
4701 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4702 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4703
4704 2. PCRE was looping on encountering an indefinitely repeated back reference to
4705 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4706 Perl does - treats the match as successful.
4707
4708 ****

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