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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 923 - (show annotations)
Tue Feb 21 13:25:05 2012 UTC (3 years, 5 months ago) by ph10
File size: 227631 byte(s)
Error occurred while calculating annotation data.
Make it possible for pcretest to select which JIT compile options are used.
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
46 Version 8.30 04-February-2012
47 -----------------------------
48
49 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
50 name is defined in ctype.h.
51
52 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
53 only in quite long subpatterns.
54
55 3. Removed the function pcre_info(), which has been obsolete and deprecated
56 since it was replaced by pcre_fullinfo() in February 2000.
57
58 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
59 match a (*MARK), and the match failed at the start of the subject, a
60 reference to memory before the start of the subject could occur. This bug
61 was introduced by fix 17 of release 8.21.
62
63 5. A reference to an unset group with zero minimum repetition was giving
64 totally wrong answers (in non-JavaScript-compatibility mode). For example,
65 /(another)?(\1?)test/ matched against "hello world test". This bug was
66 introduced in release 8.13.
67
68 6. Add support for 16-bit character strings (a large amount of work involving
69 many changes and refactorings).
70
71 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
72 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
73 from a file.
74
75 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
76 rounding is not applied in this particular case).
77
78 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
79 if they appear, or are escaped, in patterns.
80
81 10. Get rid of a number of -Wunused-but-set-variable warnings.
82
83 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
84 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
85 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
86 also returns the mark "x". This bug applied to capturing parentheses,
87 non-capturing parentheses, and atomic parentheses. It also applied to some
88 assertions.
89
90 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
91 information out of configure.ac instead of relying on pcre.h.generic, which
92 is not stored in the repository.
93
94 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
95 -lreadline.
96
97 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
98
99 15. Applied Graycode's patch to put the top-level frame on the stack rather
100 than the heap when not using the stack for recursion. This gives a
101 performance improvement in many cases when recursion is not deep.
102
103 16. Experimental code added to "pcretest -C" to output the stack frame size.
104
105
106 Version 8.21 12-Dec-2011
107 ------------------------
108
109 1. Updating the JIT compiler.
110
111 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
112 are added as well.
113
114 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
115 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
116 calling _pcre_jit_exec. Some extra comments are added.
117
118 4. (*MARK) settings inside atomic groups that do not contain any capturing
119 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
120 was introduced by change 18 for 8.20.
121
122 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
123 ECMA-262 standard.
124
125 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
126 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
127 This bug was probably introduced by change 9 of 8.13.
128
129 7. While fixing 6 above, I noticed that a number of other items were being
130 incorrectly rejected as "not fixed length". This arose partly because newer
131 opcodes had not been added to the fixed-length checking code. I have (a)
132 corrected the bug and added tests for these items, and (b) arranged for an
133 error to occur if an unknown opcode is encountered while checking for fixed
134 length instead of just assuming "not fixed length". The items that were
135 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
136 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
137 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
138
139 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
140 being incorrectly compiled and would have given unpredicatble results.
141
142 9. A possessively repeated subpattern with minimum repeat count greater than
143 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
144 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
145 the first (A) could occur when it should not.
146
147 10. Add a cast and remove a redundant test from the code.
148
149 11. JIT should use pcre_malloc/pcre_free for allocation.
150
151 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
152 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
153 is anything other than /usr, -L is still shown).
154
155 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
156
157 14. Perl does not support \N without a following name in a [] class; PCRE now
158 also gives an error.
159
160 15. If a forward reference was repeated with an upper limit of around 2000,
161 it caused the error "internal error: overran compiling workspace". The
162 maximum number of forward references (including repeats) was limited by the
163 internal workspace, and dependent on the LINK_SIZE. The code has been
164 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
165 the default depends on LINK_SIZE. There is a new upper limit (for safety)
166 of around 200,000 forward references. While doing this, I also speeded up
167 the filling in of repeated forward references.
168
169 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
170 incorrectly expecting the subject to contain another "a" after the start.
171
172 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
173 in the match, the SKIP should be ignored. This was not happening; instead
174 the SKIP was being treated as NOMATCH. For patterns such as
175 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
176 tested.
177
178 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
179 now much more compatible with Perl, in particular in cases where the result
180 is a non-match for a non-anchored pattern. For example, if
181 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
182 "m", where previously it did not return a name. A side effect of this
183 change is that for partial matches, the last encountered mark name is
184 returned, as for non matches. A number of tests that were previously not
185 Perl-compatible have been moved into the Perl-compatible test files. The
186 refactoring has had the pleasing side effect of removing one argument from
187 the match() function, thus reducing its stack requirements.
188
189 19. If the /S+ option was used in pcretest to study a pattern using JIT,
190 subsequent uses of /S (without +) incorrectly behaved like /S+.
191
192 21. Retrieve executable code size support for the JIT compiler and fixing
193 some warnings.
194
195 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
196 not work when the shorter character appeared right at the end of the
197 subject string.
198
199 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
200 systems.
201
202 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
203 output it when the /M option is used in pcretest.
204
205 25. The CheckMan script was not being included in the distribution. Also, added
206 an explicit "perl" to run Perl scripts from the PrepareRelease script
207 because this is reportedly needed in Windows.
208
209 26. If study data was being save in a file and studying had not found a set of
210 "starts with" bytes for the pattern, the data written to the file (though
211 never used) was taken from uninitialized memory and so caused valgrind to
212 complain.
213
214 27. Updated RunTest.bat as provided by Sheri Pierce.
215
216 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
217
218 29. Computation of memory usage for the table of capturing group names was
219 giving an unnecessarily large value.
220
221
222 Version 8.20 21-Oct-2011
223 ------------------------
224
225 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
226 a POSIX class. After further experiments with Perl, which convinced me that
227 Perl has bugs and confusions, a closing square bracket is no longer allowed
228 in a POSIX name. This bug also affected patterns with classes that started
229 with full stops.
230
231 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
232 captured substring, but while checking the failing first alternative,
233 substring 1 is temporarily captured. If the output vector supplied to
234 pcre_exec() was not big enough for this capture, the yield of the function
235 was still zero ("insufficient space for captured substrings"). This cannot
236 be totally fixed without adding another stack variable, which seems a lot
237 of expense for a edge case. However, I have improved the situation in cases
238 such as /(a)(b)x|abc/ matched against "abc", where the return code
239 indicates that fewer than the maximum number of slots in the ovector have
240 been set.
241
242 3. Related to (2) above: when there are more back references in a pattern than
243 slots in the output vector, pcre_exec() uses temporary memory during
244 matching, and copies in the captures as far as possible afterwards. It was
245 using the entire output vector, but this conflicts with the specification
246 that only 2/3 is used for passing back captured substrings. Now it uses
247 only the first 2/3, for compatibility. This is, of course, another edge
248 case.
249
250 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
251 main code base, and can be used by building with --enable-jit. When this is
252 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
253 runtime --no-jit option is given.
254
255 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
256 ovector, the return from the function was zero, implying that there were
257 other matches that did not fit. The correct "exactly full" value is now
258 returned.
259
260 6. If a subpattern that was called recursively or as a subroutine contained
261 (*PRUNE) or any other control that caused it to give a non-standard return,
262 invalid errors such as "Error -26 (nested recursion at the same subject
263 position)" or even infinite loops could occur.
264
265 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
266 computing the minimum length on reaching *ACCEPT, and so ended up with the
267 wrong value of 1 rather than 0. Further investigation indicates that
268 computing a minimum subject length in the presence of *ACCEPT is difficult
269 (think back references, subroutine calls), and so I have changed the code
270 so that no minimum is registered for a pattern that contains *ACCEPT.
271
272 8. If (*THEN) was present in the first (true) branch of a conditional group,
273 it was not handled as intended. [But see 16 below.]
274
275 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
276 Sheri Pierce.
277
278 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
279 the first byte in a match must be "a".
280
281 11. Change 17 for 8.13 increased the recursion depth for patterns like
282 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
283 pattern contains any instances of (*THEN). If it does not, the old
284 optimizations are restored. It would be nice to do this on a per-group
285 basis, but at the moment that is not feasible.
286
287 12. In some environments, the output of pcretest -C is CRLF terminated. This
288 broke RunTest's code that checks for the link size. A single white space
289 character after the value is now allowed for.
290
291 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
292 For "fr", it uses the Windows-specific input and output files.
293
294 14. If (*THEN) appeared in a group that was called recursively or as a
295 subroutine, it did not work as intended. [But see next item.]
296
297 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
298 pattern fragments (but not containing any | characters). If A and B are
299 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
300 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
301 D. In other words, Perl considers parentheses that do not contain any |
302 characters to be part of a surrounding alternative, whereas PCRE was
303 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
304 differently. PCRE now behaves in the same way as Perl, except in the case
305 of subroutine/recursion calls such as (?1) which have in any case always
306 been different (but PCRE had them first :-).
307
308 16. Related to 15 above: Perl does not treat the | in a conditional group as
309 creating alternatives. Such a group is treated in the same way as an
310 ordinary group without any | characters when processing (*THEN). PCRE has
311 been changed to match Perl's behaviour.
312
313 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
314 RunGrepTest script failed.
315
316 18. Change 22 for version 13 caused atomic groups to use more stack. This is
317 inevitable for groups that contain captures, but it can lead to a lot of
318 stack use in large patterns. The old behaviour has been restored for atomic
319 groups that do not contain any capturing parentheses.
320
321 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
322 suppress the check for a minimum subject length at run time. (If it was
323 given to pcre_exec() or pcre_dfa_exec() it did work.)
324
325 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
326 fail to work when decoding hex characters in data strings in EBCDIC
327 environments.
328
329 21. It appears that in at least one Mac OS environment, the isxdigit() function
330 is implemented as a macro that evaluates to its argument more than once,
331 contravening the C 90 Standard (I haven't checked a later standard). There
332 was an instance in pcretest which caused it to go wrong when processing
333 \x{...} escapes in subject strings. The has been rewritten to avoid using
334 things like p++ in the argument of isxdigit().
335
336
337 Version 8.13 16-Aug-2011
338 ------------------------
339
340 1. The Unicode data tables have been updated to Unicode 6.0.0.
341
342 2. Two minor typos in pcre_internal.h have been fixed.
343
344 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
345 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
346 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
347
348 4. There were a number of related bugs in the code for matching backrefences
349 caselessly in UTF-8 mode when codes for the characters concerned were
350 different numbers of bytes. For example, U+023A and U+2C65 are an upper
351 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
352 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
353 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
354 2-byte code at the end of the subject (it thought there wasn't enough data
355 left).
356
357 5. Comprehensive information about what went wrong is now returned by
358 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
359 as the output vector has at least 2 elements. The offset of the start of
360 the failing character and a reason code are placed in the vector.
361
362 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
363 now returned is for the first byte of the failing character, instead of the
364 last byte inspected. This is an incompatible change, but I hope it is small
365 enough not to be a problem. It makes the returned offset consistent with
366 pcre_exec() and pcre_dfa_exec().
367
368 7. pcretest now gives a text phrase as well as the error number when
369 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
370 failure, the offset and reason code are output.
371
372 8. When \R was used with a maximizing quantifier it failed to skip backwards
373 over a \r\n pair if the subsequent match failed. Instead, it just skipped
374 back over a single character (\n). This seems wrong (because it treated the
375 two characters as a single entity when going forwards), conflicts with the
376 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
377 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
378 has been changed.
379
380 9. Some internal refactoring has changed the processing so that the handling
381 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
382 time (the PCRE_DOTALL option was changed this way some time ago: version
383 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
384 which was always a bit of a fudge. It also means that there is one less
385 argument for the match() function, which reduces its stack requirements
386 slightly. This change also fixes an incompatibility with Perl: the pattern
387 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
388
389 10. More internal refactoring has drastically reduced the number of recursive
390 calls to match() for possessively repeated groups such as (abc)++ when
391 using pcre_exec().
392
393 11. While implementing 10, a number of bugs in the handling of groups were
394 discovered and fixed:
395
396 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
397 (a|)*(?1) gave a compile-time internal error.
398 ((a|)+)+ did not notice that the outer group could match an empty string.
399 (^a|^)+ was not marked as anchored.
400 (.*a|.*)+ was not marked as matching at start or after a newline.
401
402 12. Yet more internal refactoring has removed another argument from the match()
403 function. Special calls to this function are now indicated by setting a
404 value in a variable in the "match data" data block.
405
406 13. Be more explicit in pcre_study() instead of relying on "default" for
407 opcodes that mean there is no starting character; this means that when new
408 ones are added and accidentally left out of pcre_study(), testing should
409 pick them up.
410
411 14. The -s option of pcretest has been documented for ages as being an old
412 synonym of -m (show memory usage). I have changed it to mean "force study
413 for every regex", that is, assume /S for every regex. This is similar to -i
414 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
415 using it. It makes it easier to run collections of tests with and without
416 study enabled, and thereby test pcre_study() more easily. All the standard
417 tests are now run with and without -s (but some patterns can be marked as
418 "never study" - see 20 below).
419
420 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
421 restoration of the capturing data to the outer values was not happening
422 correctly.
423
424 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
425 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
426 pattern had matched an empty string, and so incorrectly returned a no
427 match.
428
429 17. There was optimizing code for the last branch of non-capturing parentheses,
430 and also for the obeyed branch of a conditional subexpression, which used
431 tail recursion to cut down on stack usage. Unfortunately, now that there is
432 the possibility of (*THEN) occurring in these branches, tail recursion is
433 no longer possible because the return has to be checked for (*THEN). These
434 two optimizations have therefore been removed. [But see 8.20/11 above.]
435
436 18. If a pattern containing \R was studied, it was assumed that \R always
437 matched two bytes, thus causing the minimum subject length to be
438 incorrectly computed because \R can also match just one byte.
439
440 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
441 was incorrectly computed.
442
443 20. If /S is present twice on a test pattern in pcretest input, it now
444 *disables* studying, thereby overriding the use of -s on the command line
445 (see 14 above). This is necessary for one or two tests to keep the output
446 identical in both cases.
447
448 21. When (*ACCEPT) was used in an assertion that matched an empty string and
449 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
450
451 22. When an atomic group that contained a capturing parenthesis was
452 successfully matched, but the branch in which it appeared failed, the
453 capturing was not being forgotten if a higher numbered group was later
454 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
455 group 1 to "a", when in fact it should be unset. This applied to multi-
456 branched capturing and non-capturing groups, repeated or not, and also to
457 positive assertions (capturing in negative assertions does not happen
458 in PCRE) and also to nested atomic groups.
459
460 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
461 subject after a captured substring, to make it easier to tell which of a
462 number of identical substrings has been captured.
463
464 24. The way atomic groups are processed by pcre_exec() has been changed so that
465 if they are repeated, backtracking one repetition now resets captured
466 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
467 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
468 "aaa". Previously, it would have been "a". As part of this code
469 refactoring, the way recursive calls are handled has also been changed.
470
471 25. If an assertion condition captured any substrings, they were not passed
472 back unless some other capturing happened later. For example, if
473 (?(?=(a))a) was matched against "a", no capturing was returned.
474
475 26. When studying a pattern that contained subroutine calls or assertions,
476 the code for finding the minimum length of a possible match was handling
477 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
478 group 1 called group 2 while simultaneously a separate group 2 called group
479 1). A stack overflow occurred in this case. I have fixed this by limiting
480 the recursion depth to 10.
481
482 27. Updated RunTest.bat in the distribution to the version supplied by Tom
483 Fortmann. This supports explicit test numbers on the command line, and has
484 argument validation and error reporting.
485
486 28. An instance of \X with an unlimited repeat could fail if at any point the
487 first character it looked at was a mark character.
488
489 29. Some minor code refactoring concerning Unicode properties and scripts
490 should reduce the stack requirement of match() slightly.
491
492 30. Added the '=' option to pcretest to check the setting of unused capturing
493 slots at the end of the pattern, which are documented as being -1, but are
494 not included in the return count.
495
496 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
497 compiled something random. Now it gives a compile-time error (as does
498 Perl).
499
500 32. A *MARK encountered during the processing of a positive assertion is now
501 recorded and passed back (compatible with Perl).
502
503 33. If --only-matching or --colour was set on a pcregrep call whose pattern
504 had alternative anchored branches, the search for a second match in a line
505 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
506 matched the line "0102" twice. The same bug affected patterns that started
507 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
508 twice.
509
510 34. Previously, PCRE did not allow quantification of assertions. However, Perl
511 does, and because of capturing effects, quantifying parenthesized
512 assertions may at times be useful. Quantifiers are now allowed for
513 parenthesized assertions.
514
515 35. A minor code tidy in pcre_compile() when checking options for \R usage.
516
517 36. \g was being checked for fancy things in a character class, when it should
518 just be a literal "g".
519
520 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
521 appearance of a nested POSIX class supersedes an apparent external class.
522 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
523 unescaped square brackets may also appear as part of class names. For
524 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
525 more like Perl. (But see 8.20/1 above.)
526
527 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
528 was because it thought it was \N{name}, which is not supported).
529
530 39. Add minix to OS list not supporting the -S option in pcretest.
531
532 40. PCRE tries to detect cases of infinite recursion at compile time, but it
533 cannot analyze patterns in sufficient detail to catch mutual recursions
534 such as ((?1))((?2)). There is now a runtime test that gives an error if a
535 subgroup is called recursively as a subpattern for a second time at the
536 same position in the subject string. In previous releases this might have
537 been caught by the recursion limit, or it might have run out of stack.
538
539 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
540 happen only once. PCRE was, however incorrectly giving a compile time error
541 "recursive call could loop indefinitely" because it cannot analyze the
542 pattern in sufficient detail. The compile time test no longer happens when
543 PCRE is compiling a conditional subpattern, but actual runaway loops are
544 now caught at runtime (see 40 above).
545
546 42. It seems that Perl allows any characters other than a closing parenthesis
547 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
548 has been changed to be the same.
549
550 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
551 as not to get warnings when autogen.sh is called. Also changed
552 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
553
554 44. To help people who use pcregrep to scan files containing exceedingly long
555 lines, the following changes have been made:
556
557 (a) The default value of the buffer size parameter has been increased from
558 8K to 20K. (The actual buffer used is three times this size.)
559
560 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
561 PCRE is built.
562
563 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
564 to be set at run time.
565
566 (d) Numerical values in pcregrep options can be followed by K or M, for
567 example --buffer-size=50K.
568
569 (e) If a line being scanned overflows pcregrep's buffer, an error is now
570 given and the return code is set to 2.
571
572 45. Add a pointer to the latest mark to the callout data block.
573
574 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
575 partial match of an empty string instead of no match. This was specific to
576 the use of ".".
577
578 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
579 complete match instead of a partial match. This bug was dependent on both
580 the PCRE_UTF8 and PCRE_DOTALL options being set.
581
582 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
583 starting byte set, because \b was not being ignored.
584
585
586 Version 8.12 15-Jan-2011
587 ------------------------
588
589 1. Fixed some typos in the markup of the man pages, and wrote a script that
590 checks for such things as part of the documentation building process.
591
592 2. On a big-endian 64-bit system, pcregrep did not correctly process the
593 --match-limit and --recursion-limit options (added for 8.11). In
594 particular, this made one of the standard tests fail. (The integer value
595 went into the wrong half of a long int.)
596
597 3. If the --colour option was given to pcregrep with -v (invert match), it
598 did strange things, either producing crazy output, or crashing. It should,
599 of course, ignore a request for colour when reporting lines that do not
600 match.
601
602 4. Another pcregrep bug caused similar problems if --colour was specified with
603 -M (multiline) and the pattern match finished with a line ending.
604
605 5. In pcregrep, when a pattern that ended with a literal newline sequence was
606 matched in multiline mode, the following line was shown as part of the
607 match. This seems wrong, so I have changed it.
608
609 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
610 the check for further matches in the same line (so they could be coloured)
611 to overrun the end of the current line. If another match was found, it was
612 incorrectly shown (and then shown again when found in the next line).
613
614 7. If pcregrep was compiled under Windows, there was a reference to the
615 function pcregrep_exit() before it was defined. I am assuming this was
616 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
617 reported by a user. I've moved the definition above the reference.
618
619
620 Version 8.11 10-Dec-2010
621 ------------------------
622
623 1. (*THEN) was not working properly if there were untried alternatives prior
624 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
625 backtracked to try for "b" instead of moving to the next alternative branch
626 at the same level (in this case, to look for "c"). The Perl documentation
627 is clear that when (*THEN) is backtracked onto, it goes to the "next
628 alternative in the innermost enclosing group".
629
630 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
631 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
632 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
633 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
634 (*THEN).
635
636 3. If \s appeared in a character class, it removed the VT character from
637 the class, even if it had been included by some previous item, for example
638 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
639 of \s, but is part of the POSIX "space" class.)
640
641 4. A partial match never returns an empty string (because you can always
642 match an empty string at the end of the subject); however the checking for
643 an empty string was starting at the "start of match" point. This has been
644 changed to the "earliest inspected character" point, because the returned
645 data for a partial match starts at this character. This means that, for
646 example, /(?<=abc)def/ gives a partial match for the subject "abc"
647 (previously it gave "no match").
648
649 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
650 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
651 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
652 has an implication that the given string is incomplete (because a partial
653 match is preferred over a full match). For this reason, these items now
654 give a partial match in this situation. [Aside: previously, the one case
655 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
656 match rather than a full match, which was wrong by the old rules, but is
657 now correct.]
658
659 6. There was a bug in the handling of #-introduced comments, recognized when
660 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
661 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
662 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
663 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
664 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
665 places in pcre_compile().
666
667 7. Related to (6) above, when pcre_compile() was skipping #-introduced
668 comments when looking ahead for named forward references to subpatterns,
669 the only newline sequence it recognized was NL. It now handles newlines
670 according to the set newline convention.
671
672 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
673 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
674 cater for a lack of strerror(). These oversights have been fixed.
675
676 9. Added --match-limit and --recursion-limit to pcregrep.
677
678 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
679
680 11. When the -o option was used, pcregrep was setting a return code of 1, even
681 when matches were found, and --line-buffered was not being honoured.
682
683 12. Added an optional parentheses number to the -o and --only-matching options
684 of pcregrep.
685
686 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
687 can match an empty string. The code to do it in pcretest and pcredemo
688 needed fixing:
689
690 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
691 only one byte after an empty string match just before CRLF (this case
692 just got forgotten; "any" and "anycrlf" were OK).
693
694 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
695 mode when an empty string match preceded an ASCII character followed by
696 a non-ASCII character. (The code for advancing by one character rather
697 than one byte was nonsense.)
698
699 (c) The pcredemo.c sample program did not have any code at all to handle
700 the cases when CRLF is a valid newline sequence.
701
702 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
703 as a starting offset was within the subject string. There is now a new
704 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
705 negative or greater than the length of the string. In order to test this,
706 pcretest is extended to allow the setting of negative starting offsets.
707
708 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
709 starting offset points to the beginning of a UTF-8 character was
710 unnecessarily clumsy. I tidied it up.
711
712 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
713 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
714
715 17. Nobody had reported that the --include_dir option, which was added in
716 release 7.7 should have been called --include-dir (hyphen, not underscore)
717 for compatibility with GNU grep. I have changed it to --include-dir, but
718 left --include_dir as an undocumented synonym, and the same for
719 --exclude-dir, though that is not available in GNU grep, at least as of
720 release 2.5.4.
721
722 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
723 characters from a string of bytes) have been redefined so as not to use
724 loops, in order to improve performance in some environments. At the same
725 time, I abstracted some of the common code into auxiliary macros to save
726 repetition (this should not affect the compiled code).
727
728 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
729 compile-time error is now given if \c is not followed by an ASCII
730 character, that is, a byte less than 128. (In EBCDIC mode, the code is
731 different, and any byte value is allowed.)
732
733 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
734 START_OPTIMIZE option, which is now allowed at compile time - but just
735 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
736 to pcregrep and other applications that have no direct access to PCRE
737 options. The new /Y option in pcretest sets this option when calling
738 pcre_compile().
739
740 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
741 back references. Groups containing recursive back references were forced to
742 be atomic by that change, but in the case of named groups, the amount of
743 memory required was incorrectly computed, leading to "Failed: internal
744 error: code overflow". This has been fixed.
745
746 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
747 pcretest.c, to avoid build problems in some Borland environments.
748
749
750 Version 8.10 25-Jun-2010
751 ------------------------
752
753 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
754 THEN.
755
756 2. (*ACCEPT) was not working when inside an atomic group.
757
758 3. Inside a character class, \B is treated as a literal by default, but
759 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
760 causes the error). The code is unchanged, but I tidied the documentation.
761
762 4. Inside a character class, PCRE always treated \R and \X as literals,
763 whereas Perl faults them if its -w option is set. I have changed PCRE so
764 that it faults them when PCRE_EXTRA is set.
765
766 5. Added support for \N, which always matches any character other than
767 newline. (It is the same as "." when PCRE_DOTALL is not set.)
768
769 6. When compiling pcregrep with newer versions of gcc which may have
770 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
771 declared with attribute warn_unused_result" were given. Just casting the
772 result to (void) does not stop the warnings; a more elaborate fudge is
773 needed. I've used a macro to implement this.
774
775 7. Minor change to pcretest.c to avoid a compiler warning.
776
777 8. Added four artifical Unicode properties to help with an option to make
778 \s etc use properties (see next item). The new properties are: Xan
779 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
780
781 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
782 use Unicode properties. (*UCP) at the start of a pattern can be used to set
783 this option. Modified pcretest to add /W to test this facility. Added
784 REG_UCP to make it available via the POSIX interface.
785
786 10. Added --line-buffered to pcregrep.
787
788 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
789 studied, and the match started with a letter with a code point greater than
790 127 whose first byte was different to the first byte of the other case of
791 the letter, the other case of this starting letter was not recognized
792 (#976).
793
794 12. If a pattern that was studied started with a repeated Unicode property
795 test, for example, \p{Nd}+, there was the theoretical possibility of
796 setting up an incorrect bitmap of starting bytes, but fortunately it could
797 not have actually happened in practice until change 8 above was made (it
798 added property types that matched character-matching opcodes).
799
800 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
801 possible starting bytes for non-anchored patterns.
802
803 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
804 \R, and also a number of cases that involve Unicode properties, both
805 explicit and implicit when PCRE_UCP is set.
806
807 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
808 input, it could crash or give wrong results if characters with values
809 greater than 0xc0 were present in the subject string. (Detail: it assumed
810 UTF-8 input when processing these items.)
811
812 16. Added a lot of (int) casts to avoid compiler warnings in systems where
813 size_t is 64-bit (#991).
814
815 17. Added a check for running out of memory when PCRE is compiled with
816 --disable-stack-for-recursion (#990).
817
818 18. If the last data line in a file for pcretest does not have a newline on
819 the end, a newline was missing in the output.
820
821 19. The default pcre_chartables.c file recognizes only ASCII characters (values
822 less than 128) in its various bitmaps. However, there is a facility for
823 generating tables according to the current locale when PCRE is compiled. It
824 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
825 characters, are recognized by isspace() and therefore were getting set in
826 these tables, and indeed these tables seem to approximate to ISO 8859. This
827 caused a problem in UTF-8 mode when pcre_study() was used to create a list
828 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
829 which of course cannot start UTF-8 characters. I have changed the code so
830 that only real ASCII characters (less than 128) and the correct starting
831 bytes for UTF-8 encodings are set for characters greater than 127 when in
832 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
833 altogether.)
834
835 20. Added the /T option to pcretest so as to be able to run tests with non-
836 standard character tables, thus making it possible to include the tests
837 used for 19 above in the standard set of tests.
838
839 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
840 reference to a subpattern the other side of a comment that contains an
841 opening parenthesis caused either an internal compiling error, or a
842 reference to the wrong subpattern.
843
844
845 Version 8.02 19-Mar-2010
846 ------------------------
847
848 1. The Unicode data tables have been updated to Unicode 5.2.0.
849
850 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
851 configured.
852
853 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
854 original author of that file, following a query about its status.
855
856 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
857 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
858
859 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
860 quantifier applied to a forward-referencing subroutine call, could compile
861 incorrect code or give the error "internal error: previously-checked
862 referenced subpattern not found".
863
864 6. Both MS Visual Studio and Symbian OS have problems with initializing
865 variables to point to external functions. For these systems, therefore,
866 pcre_malloc etc. are now initialized to local functions that call the
867 relevant global functions.
868
869 7. There were two entries missing in the vectors called coptable and poptable
870 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
871 I've fixed the data, and added a kludgy way of testing at compile time that
872 the lengths are correct (equal to the number of opcodes).
873
874 8. Following on from 7, I added a similar kludge to check the length of the
875 eint vector in pcreposix.c.
876
877 9. Error texts for pcre_compile() are held as one long string to avoid too
878 much relocation at load time. To find a text, the string is searched,
879 counting zeros. There was no check for running off the end of the string,
880 which could happen if a new error number was added without updating the
881 string.
882
883 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
884
885 11. \K was not working if it appeared in an atomic group or in a group that
886 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
887 \K is "not well defined" if used in an assertion. PCRE now accepts it if
888 the assertion is positive, but not if it is negative.
889
890 12. Change 11 fortuitously reduced the size of the stack frame used in the
891 "match()" function of pcre_exec.c by one pointer. Forthcoming
892 implementation of support for (*MARK) will need an extra pointer on the
893 stack; I have reserved it now, so that the stack frame size does not
894 decrease.
895
896 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
897 item in branch that calls a recursion is a subroutine call - as in the
898 second branch in the above example - was incorrectly given the compile-
899 time error "recursive call could loop indefinitely" because pcre_compile()
900 was not correctly checking the subroutine for matching a non-empty string.
901
902 14. The checks for overrunning compiling workspace could trigger after an
903 overrun had occurred. This is a "should never occur" error, but it can be
904 triggered by pathological patterns such as hundreds of nested parentheses.
905 The checks now trigger 100 bytes before the end of the workspace.
906
907 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
908
909
910 Version 8.01 19-Jan-2010
911 ------------------------
912
913 1. If a pattern contained a conditional subpattern with only one branch (in
914 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
915 computed the wrong minimum data length (which is of course zero for such
916 subpatterns). This could cause incorrect "no match" results.
917
918 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
919 the pattern is reset in the first branch, pcre_compile() failed with
920 "internal error: code overflow at offset...". This happened only when
921 the reset was to the original external option setting. (An optimization
922 abstracts leading options settings into an external setting, which was the
923 cause of this.)
924
925 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
926 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
927 assertion pattern did not match (meaning that the assertion was true), it
928 was incorrectly treated as false if the SKIP had been reached during the
929 matching. This also applied to assertions used as conditions.
930
931 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
932 assertion subpattern, including such a pattern used as a condition,
933 unpredictable results occurred, instead of the error return
934 PCRE_ERROR_DFA_UITEM.
935
936 5. The C++ GlobalReplace function was not working like Perl for the special
937 situation when an empty string is matched. It now does the fancy magic
938 stuff that is necessary.
939
940 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
941 removed. (These were left over from very, very early versions of PCRE.)
942
943 7. Some cosmetic changes to the code to make life easier when compiling it
944 as part of something else:
945
946 (a) Change DEBUG to PCRE_DEBUG.
947
948 (b) In pcre_compile(), rename the member of the "branch_chain" structure
949 called "current" as "current_branch", to prevent a collision with the
950 Linux macro when compiled as a kernel module.
951
952 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
953 prevent a collision with the Linux macro when compiled as a kernel
954 module.
955
956 8. In pcre_compile() there are some checks for integer overflows that used to
957 cast potentially large values to (double). This has been changed to that
958 when building, a check for int64_t is made, and if it is found, it is used
959 instead, thus avoiding the use of floating point arithmetic. (There is no
960 other use of FP in PCRE.) If int64_t is not found, the fallback is to
961 double.
962
963 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
964 2005 (difference between two addresses compared to an unsigned value).
965
966 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
967 custom one, because of the following reported problem in Windows:
968
969 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
970 under Win32.
971 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
972 therefore missing the function definition.
973 - The compiler thus generates a "C" signature for the test function.
974 - The linker fails to find the "C" function.
975 - PCRE fails to configure if asked to do so against libbz2.
976
977 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
978 messages were output:
979
980 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
981 rerunning libtoolize, to keep the correct libtool macros in-tree.
982 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
983
984 I have done both of these things.
985
986 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
987 most of the time, it *can* run out if it is given a pattern that contains a
988 runaway infinite recursion. I updated the discussion in the pcrestack man
989 page.
990
991 13. Now that we have gone to the x.xx style of version numbers, the minor
992 version may start with zero. Using 08 or 09 is a bad idea because users
993 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
994 interpreted as invalid octal numbers. I've updated the previous comment in
995 configure.ac, and also added a check that gives an error if 08 or 09 are
996 used.
997
998 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
999 causing partial matching to fail when the end of the subject matched \W
1000 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1001
1002 15. There were some discrepancies between the declarations in pcre_internal.h
1003 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1004 their definitions. The declarations used "const uschar *" and the
1005 definitions used USPTR. Even though USPTR is normally defined as "const
1006 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1007 reported that: "This difference in casting confuses some C++ compilers, for
1008 example, SunCC recognizes above declarations as different functions and
1009 generates broken code for hbpcre." I have changed the declarations to use
1010 USPTR.
1011
1012 16. GNU libtool is named differently on some systems. The autogen.sh script now
1013 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1014 (FreeBSD).
1015
1016 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1017 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1018 comment: "Figure out how to create a longlong from a string: strtoll and
1019 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1020 instance, but it only takes 2 args instead of 3!"
1021
1022 18. A subtle bug concerned with back references has been fixed by a change of
1023 specification, with a corresponding code fix. A pattern such as
1024 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1025 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1026 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1027 same bug. Such groups have to be quantified to be useful, or contained
1028 inside another quantified group. (If there's no repetition, the reference
1029 can never match.) The problem arises because, having left the group and
1030 moved on to the rest of the pattern, a later failure that backtracks into
1031 the group uses the captured value from the final iteration of the group
1032 rather than the correct earlier one. I have fixed this in PCRE by forcing
1033 any group that contains a reference to itself to be an atomic group; that
1034 is, there cannot be any backtracking into it once it has completed. This is
1035 similar to recursive and subroutine calls.
1036
1037
1038 Version 8.00 19-Oct-09
1039 ----------------------
1040
1041 1. The table for translating pcre_compile() error codes into POSIX error codes
1042 was out-of-date, and there was no check on the pcre_compile() error code
1043 being within the table. This could lead to an OK return being given in
1044 error.
1045
1046 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1047 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1048 in a Windows environment.
1049
1050 3. The pcregrep --count option prints the count for each file even when it is
1051 zero, as does GNU grep. However, pcregrep was also printing all files when
1052 --files-with-matches was added. Now, when both options are given, it prints
1053 counts only for those files that have at least one match. (GNU grep just
1054 prints the file name in this circumstance, but including the count seems
1055 more useful - otherwise, why use --count?) Also ensured that the
1056 combination -clh just lists non-zero counts, with no names.
1057
1058 4. The long form of the pcregrep -F option was incorrectly implemented as
1059 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1060 but it seems right to fix it, and I didn't think it was worth preserving
1061 the old behaviour.
1062
1063 5. The command line items --regex=pattern and --regexp=pattern were not
1064 recognized by pcregrep, which required --regex pattern or --regexp pattern
1065 (with a space rather than an '='). The man page documented the '=' forms,
1066 which are compatible with GNU grep; these now work.
1067
1068 6. No libpcreposix.pc file was created for pkg-config; there was just
1069 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1070
1071 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1072 when UCP support is not needed, by modifying the Python script that
1073 generates it from Unicode data files. This should not matter if the module
1074 is correctly used as a library, but I received one complaint about 50K of
1075 unwanted data. My guess is that the person linked everything into his
1076 program rather than using a library. Anyway, it does no harm.
1077
1078 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1079 was a minimum greater than 1 for a wide character in a possessive
1080 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1081 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1082 character. Chaos in the form of incorrect output or a compiling loop could
1083 result.
1084
1085 9. The restrictions on what a pattern can contain when partial matching is
1086 requested for pcre_exec() have been removed. All patterns can now be
1087 partially matched by this function. In addition, if there are at least two
1088 slots in the offset vector, the offset of the earliest inspected character
1089 for the match and the offset of the end of the subject are set in them when
1090 PCRE_ERROR_PARTIAL is returned.
1091
1092 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1093 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1094 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1095 and may be more useful for multi-segment matching.
1096
1097 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1098 used to be given if ever the end of the subject was reached; now it is
1099 given only if matching could not proceed because another character was
1100 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1101 string "Z", which now yields "no match" instead of "partial match". In the
1102 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1103 final character ended with (*FAIL).
1104
1105 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1106 if the pattern had a "must contain" character that was already found in the
1107 earlier partial match, unless partial matching was again requested. For
1108 example, with the pattern /dog.(body)?/, the "must contain" character is
1109 "g". If the first part-match was for the string "dog", restarting with
1110 "sbody" failed. This bug has been fixed.
1111
1112 13. The string returned by pcre_dfa_exec() after a partial match has been
1113 changed so that it starts at the first inspected character rather than the
1114 first character of the match. This makes a difference only if the pattern
1115 starts with a lookbehind assertion or \b or \B (\K is not supported by
1116 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1117 matching functions compatible, and I think it's the right thing to do.
1118
1119 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1120 so that the demonstration program is easily available in environments where
1121 PCRE has not been installed from source.
1122
1123 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1124 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1125 library.
1126
1127 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1128 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1129 is not the first non-POSIX option to be added. Clearly some people find
1130 these options useful.
1131
1132 17. If a caller to the POSIX matching function regexec() passes a non-zero
1133 value for nmatch with a NULL value for pmatch, the value of
1134 nmatch is forced to zero.
1135
1136 18. RunGrepTest did not have a test for the availability of the -u option of
1137 the diff command, as RunTest does. It now checks in the same way as
1138 RunTest, and also checks for the -b option.
1139
1140 19. If an odd number of negated classes containing just a single character
1141 interposed, within parentheses, between a forward reference to a named
1142 subpattern and the definition of the subpattern, compilation crashed with
1143 an internal error, complaining that it could not find the referenced
1144 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1145 [The bug was that it was starting one character too far in when skipping
1146 over the character class, thus treating the ] as data rather than
1147 terminating the class. This meant it could skip too much.]
1148
1149 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1150 /g option in pcretest when the pattern contains \K, which makes it possible
1151 to have an empty string match not at the start, even when the pattern is
1152 anchored. Updated pcretest and pcredemo to use this option.
1153
1154 21. If the maximum number of capturing subpatterns in a recursion was greater
1155 than the maximum at the outer level, the higher number was returned, but
1156 with unset values at the outer level. The correct (outer level) value is
1157 now given.
1158
1159 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1160 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1161 make it do so. The string so far is captured, making this feature
1162 compatible with Perl.
1163
1164 23. The tests have been re-organized, adding tests 11 and 12, to make it
1165 possible to check the Perl 5.10 features against Perl 5.10.
1166
1167 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1168 pattern matches a fixed length string. PCRE did not allow this; now it
1169 does. Neither allows recursion.
1170
1171 25. I finally figured out how to implement a request to provide the minimum
1172 length of subject string that was needed in order to match a given pattern.
1173 (It was back references and recursion that I had previously got hung up
1174 on.) This code has now been added to pcre_study(); it finds a lower bound
1175 to the length of subject needed. It is not necessarily the greatest lower
1176 bound, but using it to avoid searching strings that are too short does give
1177 some useful speed-ups. The value is available to calling programs via
1178 pcre_fullinfo().
1179
1180 26. While implementing 25, I discovered to my embarrassment that pcretest had
1181 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1182 study optimizations had never been tested with that matching function.
1183 Oops. What is worse, even when it was passed study data, there was a bug in
1184 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1185 were also very few tests of studied patterns with pcre_dfa_exec().
1186
1187 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1188 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1189 on the other side of the coin, they are no longer allowed to have different
1190 names, because these cannot be distinguished in PCRE, and this has caused
1191 confusion. (This is a difference from Perl.)
1192
1193 28. When duplicate subpattern names are present (necessarily with different
1194 numbers, as required by 27 above), and a test is made by name in a
1195 conditional pattern, either for a subpattern having been matched, or for
1196 recursion in such a pattern, all the associated numbered subpatterns are
1197 tested, and the overall condition is true if the condition is true for any
1198 one of them. This is the way Perl works, and is also more like the way
1199 testing by number works.
1200
1201
1202 Version 7.9 11-Apr-09
1203 ---------------------
1204
1205 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1206 (pcretest), all targets were linked against these libraries. This included
1207 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1208 libraries. This caused unwanted dependencies to be created. This problem
1209 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1210 pcretest is linked with readline.
1211
1212 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1213 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1214 moved outside it again, because FALSE and TRUE are already defined in AIX,
1215 but BOOL is not.
1216
1217 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1218 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1219
1220 4. The pcregrep documentation said spaces were inserted as well as colons (or
1221 hyphens) following file names and line numbers when outputting matching
1222 lines. This is not true; no spaces are inserted. I have also clarified the
1223 wording for the --colour (or --color) option.
1224
1225 5. In pcregrep, when --colour was used with -o, the list of matching strings
1226 was not coloured; this is different to GNU grep, so I have changed it to be
1227 the same.
1228
1229 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1230 each matching line was coloured. Now it goes on to look for further matches
1231 of any of the test patterns, which is the same behaviour as GNU grep.
1232
1233 7. A pattern that could match an empty string could cause pcregrep to loop; it
1234 doesn't make sense to accept an empty string match in pcregrep, so I have
1235 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1236 seems to be how GNU grep behaves.
1237
1238 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1239 start or after a newline", because the conditional assertion was not being
1240 correctly handled. The rule now is that both the assertion and what follows
1241 in the first alternative must satisfy the test.
1242
1243 9. If auto-callout was enabled in a pattern with a conditional group whose
1244 condition was an assertion, PCRE could crash during matching, both with
1245 pcre_exec() and pcre_dfa_exec().
1246
1247 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1248 used for matching.
1249
1250 11. Unicode property support in character classes was not working for
1251 characters (bytes) greater than 127 when not in UTF-8 mode.
1252
1253 12. Added the -M command line option to pcretest.
1254
1255 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1256
1257 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1258
1259 16. Added comments and documentation about mis-use of no_arg in the C++
1260 wrapper.
1261
1262 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1263 from Martin Jerabek that uses macro names for all relevant character and
1264 string constants.
1265
1266 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1267 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1268 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1269 these, but not everybody uses configure.
1270
1271 19. A conditional group that had only one branch was not being correctly
1272 recognized as an item that could match an empty string. This meant that an
1273 enclosing group might also not be so recognized, causing infinite looping
1274 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1275 with the subject "ab", where knowledge that the repeated group can match
1276 nothing is needed in order to break the loop.
1277
1278 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1279 exec(), but without supplying a callout function, matching went wrong.
1280
1281 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1282 leak if the size of the offset vector was greater than 30. When the vector
1283 is smaller, the saved offsets during recursion go onto a local stack
1284 vector, but for larger vectors malloc() is used. It was failing to free
1285 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1286 error, in fact).
1287
1288 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1289 heapframe that is used only when UTF-8 support is enabled. This caused no
1290 problem, but was untidy.
1291
1292 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1293 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1294 included within another project.
1295
1296 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1297 slightly modified by me:
1298
1299 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1300 not building pcregrep.
1301
1302 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1303 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1304
1305 25. Forward references, both numeric and by name, in patterns that made use of
1306 duplicate group numbers, could behave incorrectly or give incorrect errors,
1307 because when scanning forward to find the reference group, PCRE was not
1308 taking into account the duplicate group numbers. A pattern such as
1309 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1310
1311 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1312 the feature of a custom pointer more persuasive (as requested by a user).
1313
1314 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1315 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1316 pre-defined.
1317
1318 28. Added support for (*UTF8) at the start of a pattern.
1319
1320 29. Arrange for flags added by the "release type" setting in CMake to be shown
1321 in the configuration summary.
1322
1323
1324 Version 7.8 05-Sep-08
1325 ---------------------
1326
1327 1. Replaced UCP searching code with optimized version as implemented for Ad
1328 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1329 stage table and inline lookup instead of a function, giving speed ups of 2
1330 to 5 times on some simple patterns that I tested. Permission was given to
1331 distribute the MultiStage2.py script that generates the tables (it's not in
1332 the tarball, but is in the Subversion repository).
1333
1334 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1335 scripts.
1336
1337 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1338 a group with a zero qualifier. The result of the study could be incorrect,
1339 or the function might crash, depending on the pattern.
1340
1341 4. Caseless matching was not working for non-ASCII characters in back
1342 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1343 It now works when Unicode Property Support is available.
1344
1345 5. In pcretest, an escape such as \x{de} in the data was always generating
1346 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1347 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1348 truncation.
1349
1350 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1351
1352 7. Added two (int) casts to pcregrep when printing the difference of two
1353 pointers, in case they are 64-bit values.
1354
1355 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1356 test 2 if it fails.
1357
1358 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1359 and a #define of that name to empty if it is not externally set. This is to
1360 allow users of MSVC to set it if necessary.
1361
1362 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1363 the convenience functions in the pcre_get.c source file.
1364
1365 11. An option change at the start of a pattern that had top-level alternatives
1366 could cause overwriting and/or a crash. This command provoked a crash in
1367 some environments:
1368
1369 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1370
1371 This potential security problem was recorded as CVE-2008-2371.
1372
1373 12. For a pattern where the match had to start at the beginning or immediately
1374 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1375 pcre_dfa_exec() could read past the end of the passed subject if there was
1376 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1377 pcretest so that it places the subject at the end of its malloc-ed buffer.
1378
1379 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1380 exec() might read past the end of the data buffer in UTF-8 mode.
1381
1382 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1383 the data contained the byte 0x85 as part of a UTF-8 character within its
1384 first line. This applied both to normal and DFA matching.
1385
1386 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1387 /^[^d]*?$/8 failed to match "abc".
1388
1389 16. Added a missing copyright notice to pcrecpp_internal.h.
1390
1391 17. Make it more clear in the documentation that values returned from
1392 pcre_exec() in ovector are byte offsets, not character counts.
1393
1394 18. Tidied a few places to stop certain compilers from issuing warnings.
1395
1396 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1397 supplied by Stefan Weber. I made a further small update for 7.8 because
1398 there is a change of source arrangements: the pcre_searchfuncs.c module is
1399 replaced by pcre_ucd.c.
1400
1401
1402 Version 7.7 07-May-08
1403 ---------------------
1404
1405 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1406 a string to a long long, pretend we don't even have a long long." This is
1407 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1408
1409 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1410 pre-7.6 versions, which defined a global no_arg variable instead of putting
1411 it in the RE class. (See also #8 below.)
1412
1413 3. Remove a line of dead code, identified by coverity and reported by Nuno
1414 Lopes.
1415
1416 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1417
1418 (1) The include/exclude patterns were being applied to the whole pathnames
1419 of files, instead of just to the final components.
1420
1421 (2) If there was more than one level of directory, the subdirectories were
1422 skipped unless they satisfied the include/exclude conditions. This is
1423 inconsistent with GNU grep (and could even be seen as contrary to the
1424 pcregrep specification - which I improved to make it absolutely clear).
1425 The action now is always to scan all levels of directory, and just
1426 apply the include/exclude patterns to regular files.
1427
1428 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1429 --exclude_dir in the tests to avoid scanning .svn directories.
1430
1431 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1432 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1433 doesn't support NULs in patterns.
1434
1435 7. Added some missing "const"s to declarations of static tables in
1436 pcre_compile.c and pcre_dfa_exec.c.
1437
1438 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1439 caused by fix #2 above. (Subsequently also a second patch to fix the
1440 first patch. And a third patch - this was a messy problem.)
1441
1442 9. Applied Craig's patch to remove the use of push_back().
1443
1444 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1445 matching function regexec().
1446
1447 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1448 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1449 references. PCRE supports relative numbers with this syntax (I don't think
1450 Oniguruma does).
1451
1452 12. Previously, a group with a zero repeat such as (...){0} was completely
1453 omitted from the compiled regex. However, this means that if the group
1454 was called as a subroutine from elsewhere in the pattern, things went wrong
1455 (an internal error was given). Such groups are now left in the compiled
1456 pattern, with a new opcode that causes them to be skipped at execution
1457 time.
1458
1459 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1460 to the way PCRE behaves:
1461
1462 (a) A lone ] character is dis-allowed (Perl treats it as data).
1463
1464 (b) A back reference to an unmatched subpattern matches an empty string
1465 (Perl fails the current match path).
1466
1467 (c) A data ] in a character class must be notated as \] because if the
1468 first data character in a class is ], it defines an empty class. (In
1469 Perl it is not possible to have an empty class.) The empty class []
1470 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1471 The negative empty class [^] matches any one character, independently
1472 of the DOTALL setting.
1473
1474 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1475 non-existent subpattern following a character class starting with ']' and
1476 containing () gave an internal compiling error instead of "reference to
1477 non-existent subpattern". Fortunately, when the pattern did exist, the
1478 compiled code was correct. (When scanning forwards to check for the
1479 existencd of the subpattern, it was treating the data ']' as terminating
1480 the class, so got the count wrong. When actually compiling, the reference
1481 was subsequently set up correctly.)
1482
1483 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1484 it was being rejected as not supported by pcre_dfa_exec(), even though
1485 other assertions are supported. I have made pcre_dfa_exec() support
1486 (*FAIL).
1487
1488 16. The implementation of 13c above involved the invention of a new opcode,
1489 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1490 cannot be changed at match time, I realized I could make a small
1491 improvement to matching performance by compiling OP_ALLANY instead of
1492 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1493 on the OP_ANY path.
1494
1495 17. Compiling pcretest on Windows with readline support failed without the
1496 following two fixes: (1) Make the unistd.h include conditional on
1497 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1498
1499 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1500 ncurses library to be included for pcretest when ReadLine support is
1501 requested, but also to allow for it to be overridden. This patch came from
1502 Daniel Bergström.
1503
1504 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1505 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1506 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1507 spotting this.
1508
1509
1510 Version 7.6 28-Jan-08
1511 ---------------------
1512
1513 1. A character class containing a very large number of characters with
1514 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1515 overflow.
1516
1517 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1518 HAVE_LONG_LONG is not defined.
1519
1520 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1521 bring them up to date and include new features. This patch includes:
1522
1523 - Fixed PH's badly added libz and libbz2 support.
1524 - Fixed a problem with static linking.
1525 - Added pcredemo. [But later removed - see 7 below.]
1526 - Fixed dftables problem and added an option.
1527 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1528 HAVE_LONG_LONG.
1529 - Added readline support for pcretest.
1530 - Added an listing of the option settings after cmake has run.
1531
1532 4. A user submitted a patch to Makefile that makes it easy to create
1533 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1534 Makefile.am that cause it to include this special target, without
1535 affecting anything else. Note that the same mingw target plus all
1536 the other distribution libraries and programs are now supported
1537 when configuring with CMake (see 6 below) instead of with
1538 Configure/Make.
1539
1540 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1541 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1542 exported in the Windows port". It has not yet been confirmed that the patch
1543 solves the problem, but it does no harm.
1544
1545 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1546 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1547 with CMake, and also correct the comment about stack recursion.
1548
1549 7. Remove the automatic building of pcredemo from the ./configure system and
1550 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1551 of a program that users should build themselves after PCRE is installed, so
1552 building it automatically is not really right. What is more, it gave
1553 trouble in some build environments.
1554
1555 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1556
1557
1558 Version 7.5 10-Jan-08
1559 ---------------------
1560
1561 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1562 values in parens when parsing an RE using the C++ wrapper."
1563
1564 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1565 Characters greater than 255 were excluded from the class instead of being
1566 included.
1567
1568 3. The same bug as (2) above applied to negated POSIX classes such as
1569 [:^space:].
1570
1571 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1572 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1573 I have changed it.
1574
1575 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1576 first named subpattern) and a construct such as (?&a) would reference the
1577 first named subpattern whose name started with "a" (in other words, the
1578 length check was missing). Both these problems are fixed. "Subpattern name
1579 expected" is now given for (?&) (a zero-length name), and this patch also
1580 makes it give the same error for \k'' (previously it complained that that
1581 was a reference to a non-existent subpattern).
1582
1583 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1584 this is right because (?- can be followed by option settings as well as by
1585 digits. I have, however, made the messages clearer.
1586
1587 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1588 than the number used in the conditional) now cause a compile-time error.
1589 This is actually not compatible with Perl, which accepts such patterns, but
1590 treats the conditional as always being FALSE (as PCRE used to), but it
1591 seems to me that giving a diagnostic is better.
1592
1593 8. Change "alphameric" to the more common word "alphanumeric" in comments
1594 and messages.
1595
1596 9. Fix two occurrences of "backslash" in comments that should have been
1597 "backspace".
1598
1599 10. Remove two redundant lines of code that can never be obeyed (their function
1600 was moved elsewhere).
1601
1602 11. The program that makes PCRE's Unicode character property table had a bug
1603 which caused it to generate incorrect table entries for sequences of
1604 characters that have the same character type, but are in different scripts.
1605 It amalgamated them into a single range, with the script of the first of
1606 them. In other words, some characters were in the wrong script. There were
1607 thirteen such cases, affecting characters in the following ranges:
1608
1609 U+002b0 - U+002c1
1610 U+0060c - U+0060d
1611 U+0061e - U+00612
1612 U+0064b - U+0065e
1613 U+0074d - U+0076d
1614 U+01800 - U+01805
1615 U+01d00 - U+01d77
1616 U+01d9b - U+01dbf
1617 U+0200b - U+0200f
1618 U+030fc - U+030fe
1619 U+03260 - U+0327f
1620 U+0fb46 - U+0fbb1
1621 U+10450 - U+1049d
1622
1623 12. The -o option (show only the matching part of a line) for pcregrep was not
1624 compatible with GNU grep in that, if there was more than one match in a
1625 line, it showed only the first of them. It now behaves in the same way as
1626 GNU grep.
1627
1628 13. If the -o and -v options were combined for pcregrep, it printed a blank
1629 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1630 does the same. The return code can be used to tell if there were any
1631 non-matching lines.
1632
1633 14. Added --file-offsets and --line-offsets to pcregrep.
1634
1635 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1636 infinitely looping recursion. The bug was that positive lookaheads were not
1637 being skipped when checking for a possible empty match (negative lookaheads
1638 and both kinds of lookbehind were skipped).
1639
1640 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1641 inclusion of <windows.h> to before rather than after the definition of
1642 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1643
1644 17. Specifying a possessive quantifier with a specific limit for a Unicode
1645 character property caused pcre_compile() to compile bad code, which led at
1646 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1647 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1648 caused the error; without that there was no problem.
1649
1650 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1651
1652 19. Added --enable-pcretest-libreadline.
1653
1654 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1655 RE::GlobalReplace(). As a result, the number of replacements returned was
1656 double what it should be. I removed one of the increments, but Craig sent a
1657 later patch that removed the other one (the right fix) and added unit tests
1658 that check the return values (which was not done before).
1659
1660 21. Several CMake things:
1661
1662 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1663 the names libpcre and libpcreposix, not just pcre and pcreposix.
1664
1665 (2) The above change means that pcretest and pcregrep are now correctly
1666 linked with the newly-built libraries, not previously installed ones.
1667
1668 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1669
1670 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1671 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1672 UTF-8 newline character). The key issue is that the pattern starts .*;
1673 this means that the match must be either at the beginning, or after a
1674 newline. The bug was in the code for advancing after a failed match and
1675 checking that the new position followed a newline. It was not taking
1676 account of UTF-8 characters correctly.
1677
1678 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1679 character classes. PCRE was not treating the sequence [:...:] as a
1680 character class unless the ... were all letters. Perl, however, seems to
1681 allow any characters between [: and :], though of course it rejects as
1682 unknown any "names" that contain non-letters, because all the known class
1683 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1684 for example, whereas PCRE did not - it did not recognize a POSIX character
1685 class. This seemed a bit dangerous, so the code has been changed to be
1686 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1687 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1688 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1689 Perl does, and where it didn't before.
1690
1691 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1692 Windows environments %n is disabled by default.
1693
1694
1695 Version 7.4 21-Sep-07
1696 ---------------------
1697
1698 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1699 means that a class such as [\s] counted as "explicit reference to CR or
1700 LF". That isn't really right - the whole point of the change was to try to
1701 help when there was an actual mention of one of the two characters. So now
1702 the change happens only if \r or \n (or a literal CR or LF) character is
1703 encountered.
1704
1705 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1706 of both had grown to the point where there were only 3 bits left.
1707 Fortunately, there was spare space in the data structure, and so I have
1708 moved the internal flags into a new 16-bit field to free up more option
1709 bits.
1710
1711 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1712 but did not set the internal JCHANGED flag - either of these is enough to
1713 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1714 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1715 start sets both bits.
1716
1717 4. Added options (at build time, compile time, exec time) to change \R from
1718 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1719
1720 5. doc/pcresyntax.html was missing from the distribution.
1721
1722 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1723 compatibility, even though it is no longer used.
1724
1725 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1726 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1727 windows.h file is present (where different names are used). [This was
1728 reversed later after testing - see 16 below.]
1729
1730 8. Changed all #include <config.h> to #include "config.h". There were also
1731 some further <pcre.h> cases that I changed to "pcre.h".
1732
1733 9. When pcregrep was used with the --colour option, it missed the line ending
1734 sequence off the lines that it output.
1735
1736 10. It was pointed out to me that arrays of string pointers cause lots of
1737 relocations when a shared library is dynamically loaded. A technique of
1738 using a single long string with a table of offsets can drastically reduce
1739 these. I have refactored PCRE in four places to do this. The result is
1740 dramatic:
1741
1742 Originally: 290
1743 After changing UCP table: 187
1744 After changing error message table: 43
1745 After changing table of "verbs" 36
1746 After changing table of Posix names 22
1747
1748 Thanks to the folks working on Gregex for glib for this insight.
1749
1750 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1751 unicode-properties was also set.
1752
1753 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1754
1755 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1756 checked only for CRLF.
1757
1758 14. Added casts to pcretest.c to avoid compiler warnings.
1759
1760 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1761
1762 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1763 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1764 entirely. This removes changes made in 7 above.
1765
1766 17. The CMake files have been updated, and there is now more information about
1767 building with CMake in the NON-UNIX-USE document.
1768
1769
1770 Version 7.3 28-Aug-07
1771 ---------------------
1772
1773 1. In the rejigging of the build system that eventually resulted in 7.1, the
1774 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1775 brackets there is not right, since it causes compilers to look for an
1776 installed pcre.h, not the version that is in the source that is being
1777 compiled (which of course may be different). I have changed it back to:
1778
1779 #include "pcre.h"
1780
1781 I have a vague recollection that the change was concerned with compiling in
1782 different directories, but in the new build system, that is taken care of
1783 by the VPATH setting the Makefile.
1784
1785 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1786 when the subject happened to end in the byte 0x85 (e.g. if the last
1787 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1788 characters but of course it shouldn't be taken as a newline when it is part
1789 of another character. The bug was that, for an unlimited repeat of . in
1790 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1791 characters when looking for a newline.
1792
1793 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1794
1795 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1796 in debug output.
1797
1798 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1799 long printing in the pcrecpp unittest when running under MinGW.
1800
1801 6. ESC_K was left out of the EBCDIC table.
1802
1803 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1804 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1805 limit also applies to "virtual nesting" when a pattern is recursive, and in
1806 this case 1000 isn't so big. I have been able to remove this limit at the
1807 expense of backing off one optimization in certain circumstances. Normally,
1808 when pcre_exec() would call its internal match() function recursively and
1809 immediately return the result unconditionally, it uses a "tail recursion"
1810 feature to save stack. However, when a subpattern that can match an empty
1811 string has an unlimited repetition quantifier, it no longer makes this
1812 optimization. That gives it a stack frame in which to save the data for
1813 checking that an empty string has been matched. Previously this was taken
1814 from the 1000-entry workspace that had been reserved. So now there is no
1815 explicit limit, but more stack is used.
1816
1817 8. Applied Daniel's patches to solve problems with the import/export magic
1818 syntax that is required for Windows, and which was going wrong for the
1819 pcreposix and pcrecpp parts of the library. These were overlooked when this
1820 problem was solved for the main library.
1821
1822 9. There were some crude static tests to avoid integer overflow when computing
1823 the size of patterns that contain repeated groups with explicit upper
1824 limits. As the maximum quantifier is 65535, the maximum group length was
1825 set at 30,000 so that the product of these two numbers did not overflow a
1826 32-bit integer. However, it turns out that people want to use groups that
1827 are longer than 30,000 bytes (though not repeat them that many times).
1828 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1829 made it possible to implement the integer overflow checks in a much more
1830 dynamic way, which I have now done. The artificial limitation on group
1831 length has been removed - we now have only the limit on the total length of
1832 the compiled pattern, which depends on the LINK_SIZE setting.
1833
1834 10. Fixed a bug in the documentation for get/copy named substring when
1835 duplicate names are permitted. If none of the named substrings are set, the
1836 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1837 empty string.
1838
1839 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1840 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1841 because the ] is interpreted as the first data character and the
1842 terminating ] is not found. PCRE has been made compatible with Perl in this
1843 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1844 cause memory overwriting.
1845
1846 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1847 string has been matched (to stop an infinite loop). It was not recognizing
1848 a conditional subpattern that could match an empty string if that
1849 subpattern was within another subpattern. For example, it looped when
1850 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1851 condition was not nested. This bug has been fixed.
1852
1853 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1854 past the start of the subject in the presence of bytes with the top bit
1855 set, for example "\x8aBCD".
1856
1857 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1858 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1859
1860 14. Optimized (?!) to (*FAIL).
1861
1862 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1863 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1864 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1865 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1866 does: it's just the validity check that is more restrictive.
1867
1868 16. Inserted checks for integer overflows during escape sequence (backslash)
1869 processing, and also fixed erroneous offset values for syntax errors during
1870 backslash processing.
1871
1872 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1873 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1874
1875 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1876 caused an overrun.
1877
1878 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1879 something other than just ASCII characters) inside a group that had an
1880 unlimited repeat caused a loop at compile time (while checking to see
1881 whether the group could match an empty string).
1882
1883 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1884 [\P{Any}] did so. (Error in the code for printing property names.)
1885
1886 21. An orphan \E inside a character class could cause a crash.
1887
1888 22. A repeated capturing bracket such as (A)? could cause a wild memory
1889 reference during compilation.
1890
1891 23. There are several functions in pcre_compile() that scan along a compiled
1892 expression for various reasons (e.g. to see if it's fixed length for look
1893 behind). There were bugs in these functions when a repeated \p or \P was
1894 present in the pattern. These operators have additional parameters compared
1895 with \d, etc, and these were not being taken into account when moving along
1896 the compiled data. Specifically:
1897
1898 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1899 length.
1900
1901 (b) An item such as \pL+ within a repeated group could cause crashes or
1902 loops.
1903
1904 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1905 "reference to non-existent subpattern" error.
1906
1907 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1908
1909 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1910 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1911
1912 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1913
1914 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1915 character were causing crashes (broken optimization).
1916
1917 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1918 \p or \P) caused a compile-time loop.
1919
1920 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1921 break. For example, the unstudied pattern [\r\n]A does not match the string
1922 "\r\nA" because change 7.0/46 below moves the current point on by two
1923 characters after failing to match at the start. However, the pattern \nA
1924 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1925 the same is true. There doesn't seem any very clean way out of this, but
1926 what I have chosen to do makes the common cases work: PCRE now takes note
1927 of whether there can be an explicit match for \r or \n anywhere in the
1928 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1929 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1930 pattern has explicit CR or LF references.
1931
1932 29. Added (*CR) etc for changing newline setting at start of pattern.
1933
1934
1935 Version 7.2 19-Jun-07
1936 ---------------------
1937
1938 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1939 which is apparently normally available under Windows.
1940
1941 2. Re-jig the pcregrep tests with different newline settings in an attempt
1942 to make them independent of the local environment's newline setting.
1943
1944 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1945
1946 4. Some of the "internals" tests were previously cut out when the link size
1947 was not 2, because the output contained actual offsets. The recent new
1948 "Z" feature of pcretest means that these can be cut out, making the tests
1949 usable with all link sizes.
1950
1951 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1952 stack recursion. This gives a massive performance boost under BSD, but just
1953 a small improvement under Linux. However, it saves one field in the frame
1954 in all cases.
1955
1956 6. Added more features from the forthcoming Perl 5.10:
1957
1958 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1959 recursion call. It refers to the nth most recently opened parentheses.
1960
1961 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1962 to be opened parentheses.
1963
1964 (c) Conditions that refer to capturing parentheses can be specified
1965 relatively, for example, (?(-2)... or (?(+3)...
1966
1967 (d) \K resets the start of the current match so that everything before
1968 is not part of it.
1969
1970 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1971
1972 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1973 reference syntax.
1974
1975 (g) (?| introduces a group in which the numbering of parentheses in each
1976 alternative starts with the same number.
1977
1978 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1979
1980 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1981 PCRE_INFO_JCHANGED.
1982
1983 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1984 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1985 for detecting groups that can match an empty string.
1986
1987 9. A pattern with a very large number of alternatives (more than several
1988 hundred) was running out of internal workspace during the pre-compile
1989 phase, where pcre_compile() figures out how much memory will be needed. A
1990 bit of new cunning has reduced the workspace needed for groups with
1991 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1992 workspace instead of running out of the 4096 that are available.
1993
1994 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1995
1996 11. Applied patch from Google to remove an optimization that didn't quite work.
1997 The report of the bug said:
1998
1999 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2000 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2001 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2002
2003 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2004 it matched the wrong number of bytes.
2005
2006
2007 Version 7.1 24-Apr-07
2008 ---------------------
2009
2010 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2011 that is more "standard", making use of automake and other Autotools. There
2012 is some re-arrangement of the files and adjustment of comments consequent
2013 on this.
2014
2015 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2016 for recursive directory scanning broke on some systems because the files
2017 are not scanned in any specific order and on different systems the order
2018 was different. A call to "sort" has been inserted into RunGrepTest for the
2019 approprate test as a short-term fix. In the longer term there may be an
2020 alternative.
2021
2022 3. I had an email from Eric Raymond about problems translating some of PCRE's
2023 man pages to HTML (despite the fact that I distribute HTML pages, some
2024 people do their own conversions for various reasons). The problems
2025 concerned the use of low-level troff macros .br and .in. I have therefore
2026 removed all such uses from the man pages (some were redundant, some could
2027 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2028 HTML has been updated to handle .nf/.fi and to complain if it encounters
2029 .br or .in.
2030
2031 4. Updated comments in configure.ac that get placed in config.h.in and also
2032 arranged for config.h to be included in the distribution, with the name
2033 config.h.generic, for the benefit of those who have to compile without
2034 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2035
2036 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2037 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2038 makevp.bat for the current PCRE, using the additional files
2039 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2040
2041 6. A Windows user reported a minor discrepancy with test 2, which turned out
2042 to be caused by a trailing space on an input line that had got lost in his
2043 copy. The trailing space was an accident, so I've just removed it.
2044
2045 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2046 that is needed.
2047
2048 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2049 as "const" (a) because they are and (b) because it helps the PHP
2050 maintainers who have recently made a script to detect big data structures
2051 in the php code that should be moved to the .rodata section. I remembered
2052 to update Builducptable as well, so it won't revert if ucptable.h is ever
2053 re-created.
2054
2055 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2056 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2057 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2058 support is not required. This saves 1.5-2K of code, which is important in
2059 some applications.
2060
2061 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2062 so as not to refer to the tables, even though these functions will never be
2063 called when UTF-8 support is disabled. Otherwise there are problems with a
2064 shared library.
2065
2066 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2067
2068 (a) It was defining its arguments as char * instead of void *.
2069
2070 (b) It was assuming that all moves were upwards in memory; this was true
2071 a long time ago when I wrote it, but is no longer the case.
2072
2073 The emulated memove() is provided for those environments that have neither
2074 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2075 is clearly not the case, as these two bugs were recently reported.
2076
2077 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2078 and Detrail to create the HTML documentation, the .txt form of the man
2079 pages, and it removes trailing spaces from listed files. It also creates
2080 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2081 case, it wraps all the #defines with #ifndefs. This script should be run
2082 before "make dist".
2083
2084 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2085 with Unicode property support.
2086
2087 (a) For a maximizing quantifier, if the two different cases of the
2088 character were of different lengths in their UTF-8 codings (there are
2089 some cases like this - I found 11), and the matching function had to
2090 back up over a mixture of the two cases, it incorrectly assumed they
2091 were both the same length.
2092
2093 (b) When PCRE was configured to use the heap rather than the stack for
2094 recursion during matching, it was not correctly preserving the data for
2095 the other case of a UTF-8 character when checking ahead for a match
2096 while processing a minimizing repeat. If the check also involved
2097 matching a wide character, but failed, corruption could cause an
2098 erroneous result when trying to check for a repeat of the original
2099 character.
2100
2101 13. Some tidying changes to the testing mechanism:
2102
2103 (a) The RunTest script now detects the internal link size and whether there
2104 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2105 values substituted by "configure". (The RunGrepTest script already did
2106 this for UTF-8.) The configure.ac script no longer substitutes the
2107 relevant variables.
2108
2109 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2110 with length and offset values. This means that the output is different
2111 for different internal link sizes. Test 2 is skipped for link sizes
2112 other than 2 because of this, bypassing the problem. Unfortunately,
2113 there was also a test in test 3 (the locale tests) that used /B and
2114 failed for link sizes other than 2. Rather than cut the whole test out,
2115 I have added a new /Z option to pcretest that replaces the length and
2116 offset values with spaces. This is now used to make test 3 independent
2117 of link size. (Test 2 will be tidied up later.)
2118
2119 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2120 segmentation fault instead of returning the appropriate error message.
2121
2122 15. In multiline mode when the newline sequence was set to "any", the pattern
2123 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2124 This doesn't seem right; it now treats the CRLF combination as the line
2125 ending, and so does not match in that case. It's only a pattern such as ^$
2126 that would hit this one: something like ^ABC$ would have failed after \r
2127 and then tried again after \r\n.
2128
2129 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2130 in an attempt to make files that differ only in their line terminators
2131 compare equal. This works on Linux.
2132
2133 17. Under certain error circumstances pcregrep might try to free random memory
2134 as it exited. This is now fixed, thanks to valgrind.
2135
2136 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2137 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2138 was because its rules for how to advance for /g after matching an empty
2139 string at the end of a line did not allow for this case. They now check for
2140 it specially.
2141
2142 20. pcretest is supposed to handle patterns and data of any length, by
2143 extending its buffers when necessary. It was getting this wrong when the
2144 buffer for a data line had to be extended.
2145
2146 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2147 CRLF as a newline sequence.
2148
2149 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2150 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2151 I have nevertheless tidied it up.
2152
2153 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2154
2155 24. Added a man page for pcre-config.
2156
2157
2158 Version 7.0 19-Dec-06
2159 ---------------------
2160
2161 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2162 moving to gcc 4.1.1.
2163
2164 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2165 sys/time.h, which is documented as needed for this function. It doesn't
2166 seem to matter on Linux, but it showed up on some releases of OS X.
2167
2168 3. It seems that there are systems where bytes whose values are greater than
2169 127 match isprint() in the "C" locale. The "C" locale should be the
2170 default when a C program starts up. In most systems, only ASCII printing
2171 characters match isprint(). This difference caused the output from pcretest
2172 to vary, making some of the tests fail. I have changed pcretest so that:
2173
2174 (a) When it is outputting text in the compiled version of a pattern, bytes
2175 other than 32-126 are always shown as hex escapes.
2176
2177 (b) When it is outputting text that is a matched part of a subject string,
2178 it does the same, unless a different locale has been set for the match
2179 (using the /L modifier). In this case, it uses isprint() to decide.
2180
2181 4. Fixed a major bug that caused incorrect computation of the amount of memory
2182 required for a compiled pattern when options that changed within the
2183 pattern affected the logic of the preliminary scan that determines the
2184 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2185 that the computed length was too small. The symptoms of this bug were
2186 either the PCRE error "internal error: code overflow" from pcre_compile(),
2187 or a glibc crash with a message such as "pcretest: free(): invalid next
2188 size (fast)". Examples of patterns that provoked this bug (shown in
2189 pcretest format) are:
2190
2191 /(?-x: )/x
2192 /(?x)(?-x: \s*#\s*)/
2193 /((?i)[\x{c0}])/8
2194 /(?i:[\x{c0}])/8
2195
2196 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2197 is now done differently.
2198
2199 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2200 wrapper classes; (b) implement a new function in the C++ scanner that is
2201 more efficient than the old way of doing things because it avoids levels of
2202 recursion in the regex matching; (c) add a paragraph to the documentation
2203 for the FullMatch() function.
2204
2205 6. The escape sequence \n was being treated as whatever was defined as
2206 "newline". Not only was this contrary to the documentation, which states
2207 that \n is character 10 (hex 0A), but it also went horribly wrong when
2208 "newline" was defined as CRLF. This has been fixed.
2209
2210 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2211 was being set to -1 for the "end of line" case (supposedly a value that no
2212 character can have). Though this value is never used (the check for end of
2213 line is "zero bytes in current character"), it caused compiler complaints.
2214 I've changed it to 0xffffffff.
2215
2216 8. In pcre_version.c, the version string was being built by a sequence of
2217 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2218 string (as it is for production releases) called a macro with an empty
2219 argument. The C standard says the result of this is undefined. The gcc
2220 compiler treats it as an empty string (which was what was wanted) but it is
2221 reported that Visual C gives an error. The source has been hacked around to
2222 avoid this problem.
2223
2224 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2225 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2226 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2227 of them did).
2228
2229 10. Originally, pcretest opened its input and output without "b"; then I was
2230 told that "b" was needed in some environments, so it was added for release
2231 5.0 to both the input and output. (It makes no difference on Unix-like
2232 systems.) Later I was told that it is wrong for the input on Windows. I've
2233 now abstracted the modes into two macros, to make it easier to fiddle with
2234 them, and removed "b" from the input mode under Windows.
2235
2236 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2237
2238 12. Added -help and --help to pcretest as an official way of being reminded
2239 of the options.
2240
2241 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2242 and pcrecpp.cc because they annoy compilers at high warning levels.
2243
2244 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2245
2246 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2247 scripts are not C programs :-) and which was not noticed because it works
2248 on Linux.
2249
2250 16. pcretest is supposed to handle any length of pattern and data line (as one
2251 line or as a continued sequence of lines) by extending its input buffer if
2252 necessary. This feature was broken for very long pattern lines, leading to
2253 a string of junk being passed to pcre_compile() if the pattern was longer
2254 than about 50K.
2255
2256 17. I have done a major re-factoring of the way pcre_compile() computes the
2257 amount of memory needed for a compiled pattern. Previously, there was code
2258 that made a preliminary scan of the pattern in order to do this. That was
2259 OK when PCRE was new, but as the facilities have expanded, it has become
2260 harder and harder to keep it in step with the real compile phase, and there
2261 have been a number of bugs (see for example, 4 above). I have now found a
2262 cunning way of running the real compile function in a "fake" mode that
2263 enables it to compute how much memory it would need, while actually only
2264 ever using a few hundred bytes of working memory and without too many
2265 tests of the mode. This should make future maintenance and development
2266 easier. A side effect of this work is that the limit of 200 on the nesting
2267 depth of parentheses has been removed (though this was never a serious
2268 limitation, I suspect). However, there is a downside: pcre_compile() now
2269 runs more slowly than before (30% or more, depending on the pattern). I
2270 hope this isn't a big issue. There is no effect on runtime performance.
2271
2272 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2273 newline (only possible for the last line of a file) and it was a
2274 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2275
2276 19. Added additional timing features to pcretest. (1) The -tm option now times
2277 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2278 separate command line item, by a number that specifies the number of
2279 repeats to use when timing. The default is 50000; this gives better
2280 precision, but takes uncomfortably long for very large patterns.
2281
2282 20. Extended pcre_study() to be more clever in cases where a branch of a
2283 subpattern has no definite first character. For example, (a*|b*)[cd] would
2284 previously give no result from pcre_study(). Now it recognizes that the
2285 first character must be a, b, c, or d.
2286
2287 21. There was an incorrect error "recursive call could loop indefinitely" if
2288 a subpattern (or the entire pattern) that was being tested for matching an
2289 empty string contained only one non-empty item after a nested subpattern.
2290 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2291 incorrectly, because the \d was being skipped in the check.
2292
2293 22. The pcretest program now has a new pattern option /B and a command line
2294 option -b, which is equivalent to adding /B to every pattern. This causes
2295 it to show the compiled bytecode, without the additional information that
2296 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2297 is the same as /B/I).
2298
2299 23. A new optimization is now able automatically to treat some sequences such
2300 as a*b as a*+b. More specifically, if something simple (such as a character
2301 or a simple class like \d) has an unlimited quantifier, and is followed by
2302 something that cannot possibly match the quantified thing, the quantifier
2303 is automatically "possessified".
2304
2305 24. A recursive reference to a subpattern whose number was greater than 39
2306 went wrong under certain circumstances in UTF-8 mode. This bug could also
2307 have affected the operation of pcre_study().
2308
2309 25. Realized that a little bit of performance could be had by replacing
2310 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2311
2312 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2313
2314 27. Possessive quantifiers such as a++ were previously implemented by turning
2315 them into atomic groups such as ($>a+). Now they have their own opcodes,
2316 which improves performance. This includes the automatically created ones
2317 from 23 above.
2318
2319 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2320 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2321 the first matched character to be a colon. This applied both to named and
2322 numbered groups.
2323
2324 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2325
2326 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2327 building PCRE on Windows easier, so I have included it in the distribution.
2328
2329 31. There is now a check in pcretest against a ridiculously large number being
2330 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2331 loop, the loop is abandoned.
2332
2333 32. Forward references to subpatterns in conditions such as (?(2)...) where
2334 subpattern 2 is defined later cause pcre_compile() to search forwards in
2335 the pattern for the relevant set of parentheses. This search went wrong
2336 when there were unescaped parentheses in a character class, parentheses
2337 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2338
2339 33. "Subroutine" calls and backreferences were previously restricted to
2340 referencing subpatterns earlier in the regex. This restriction has now
2341 been removed.
2342
2343 34. Added a number of extra features that are going to be in Perl 5.10. On the
2344 whole, these are just syntactic alternatives for features that PCRE had
2345 previously implemented using the Python syntax or my own invention. The
2346 other formats are all retained for compatibility.
2347
2348 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2349 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2350 also .NET compatible.
2351
2352 (b) A recursion or subroutine call to a named group can now be defined as
2353 (?&name) as well as (?P>name).
2354
2355 (c) A backreference to a named group can now be defined as \k<name> or
2356 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2357 5.10, are also .NET compatible.
2358
2359 (d) A conditional reference to a named group can now use the syntax
2360 (?(<name>) or (?('name') as well as (?(name).
2361
2362 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2363 groups (named and numbered) that are never evaluated inline, but can be
2364 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2365 is always false. There may be only one alternative in such a group.
2366
2367 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2368 as the simple (?(R). The condition is true only if the most recent
2369 recursion is that of the given number or name. It does not search out
2370 through the entire recursion stack.
2371
2372 (g) The escape \gN or \g{N} has been added, where N is a positive or
2373 negative number, specifying an absolute or relative reference.
2374
2375 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2376 some "unreachable code" warnings.
2377
2378 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2379 things, this adds five new scripts.
2380
2381 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2382 There were also incompatibilities regarding the handling of \Q..\E inside
2383 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2384 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2385
2386 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2387 matches an empty string, and forcibly breaks the loop. There were bugs in
2388 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2389 against aaaa the result was just "a" rather than "aaaa", for example. Two
2390 separate and independent bugs (that affected different cases) have been
2391 fixed.
2392
2393 39. Refactored the code to abolish the use of different opcodes for small
2394 capturing bracket numbers. This is a tidy that I avoided doing when I
2395 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2396 The new approach is not only tidier, it makes it possible to reduce the
2397 memory needed to fix the previous bug (38).
2398
2399 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2400 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2401 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2402 mode.
2403
2404 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2405 report.
2406
2407 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2408 copy construction and assignment in the C++ wrapper.
2409
2410 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2411 couple of bugs that could have given wrong results in the "--newline=crlf"
2412 case.
2413
2414 44. Added a number of casts and did some reorganization of signed/unsigned int
2415 variables following suggestions from Dair Grant. Also renamed the variable
2416 "this" as "item" because it is a C++ keyword.
2417
2418 45. Arranged for dftables to add
2419
2420 #include "pcre_internal.h"
2421
2422 to pcre_chartables.c because without it, gcc 4.x may remove the array
2423 definition from the final binary if PCRE is built into a static library and
2424 dead code stripping is activated.
2425
2426 46. For an unanchored pattern, if a match attempt fails at the start of a
2427 newline sequence, and the newline setting is CRLF or ANY, and the next two
2428 characters are CRLF, advance by two characters instead of one.
2429
2430
2431 Version 6.7 04-Jul-06
2432 ---------------------
2433
2434 1. In order to handle tests when input lines are enormously long, pcretest has
2435 been re-factored so that it automatically extends its buffers when
2436 necessary. The code is crude, but this _is_ just a test program. The
2437 default size has been increased from 32K to 50K.
2438
2439 2. The code in pcre_study() was using the value of the re argument before
2440 testing it for NULL. (Of course, in any sensible call of the function, it
2441 won't be NULL.)
2442
2443 3. The memmove() emulation function in pcre_internal.h, which is used on
2444 systems that lack both memmove() and bcopy() - that is, hardly ever -
2445 was missing a "static" storage class specifier.
2446
2447 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2448 containing an extended class (one that cannot be represented by a bitmap
2449 because it contains high-valued characters or Unicode property items, e.g.
2450 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2451 pattern, but PCRE should not loop if you do not (it no longer does).
2452 [Detail: two cases were found: (a) a repeated subpattern containing an
2453 extended class; (b) a recursive reference to a subpattern that followed a
2454 previous extended class. It wasn't skipping over the extended class
2455 correctly when UTF-8 mode was not set.]
2456
2457 5. A negated single-character class was not being recognized as fixed-length
2458 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2459 compile error "lookbehind assertion is not fixed length".
2460
2461 6. The RunPerlTest auxiliary script was showing an unexpected difference
2462 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2463 write a Perl script that can interpret lines of an input file either as
2464 byte characters or as UTF-8, which is what "perltest" was being required to
2465 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2466 can't do is switch easily at run time between having the "use utf8;" pragma
2467 or not. In the end, I fudged it by using the RunPerlTest script to insert
2468 "use utf8;" explicitly for the UTF-8 tests.
2469
2470 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2471 the end of the subject string, contrary to the documentation and to what
2472 Perl does. This was true of both matching functions. Now it matches only at
2473 the start of the subject and immediately after *internal* newlines.
2474
2475 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2476 a pointer to an int instead of a pointer to an unsigned long int. This
2477 caused problems on 64-bit systems.
2478
2479 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2480 instance of the 'standard' template library not being so standard".
2481
2482 10. There was no check on the number of named subpatterns nor the maximum
2483 length of a subpattern name. The product of these values is used to compute
2484 the size of the memory block for a compiled pattern. By supplying a very
2485 long subpattern name and a large number of named subpatterns, the size
2486 computation could be caused to overflow. This is now prevented by limiting
2487 the length of names to 32 characters, and the number of named subpatterns
2488 to 10,000.
2489
2490 11. Subpatterns that are repeated with specific counts have to be replicated in
2491 the compiled pattern. The size of memory for this was computed from the
2492 length of the subpattern and the repeat count. The latter is limited to
2493 65535, but there was no limit on the former, meaning that integer overflow
2494 could in principle occur. The compiled length of a repeated subpattern is
2495 now limited to 30,000 bytes in order to prevent this.
2496
2497 12. Added the optional facility to have named substrings with the same name.
2498
2499 13. Added the ability to use a named substring as a condition, using the
2500 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2501 are numbers (not recommended). Forward references are permitted.
2502
2503 14. Added forward references in named backreferences (if you see what I mean).
2504
2505 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2506 pattern could run off the end of the subject. For example, the pattern
2507 "(?s)(.{1,5})"8 did this with the subject "ab".
2508
2509 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2510 PCRE_CASELESS was set when matching characters that were quantified with ?
2511 or *.
2512
2513 17. A character class other than a single negated character that had a minimum
2514 but no maximum quantifier - for example [ab]{6,} - was not handled
2515 correctly by pce_dfa_exec(). It would match only one character.
2516
2517 18. A valid (though odd) pattern that looked like a POSIX character
2518 class but used an invalid character after [ (for example [[,abc,]]) caused
2519 pcre_compile() to give the error "Failed: internal error: code overflow" or
2520 in some cases to crash with a glibc free() error. This could even happen if
2521 the pattern terminated after [[ but there just happened to be a sequence of
2522 letters, a binary zero, and a closing ] in the memory that followed.
2523
2524 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2525 over the years. Originally (before any Unicode support), just the bottom 8
2526 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2527 output from "man perlunicode" includes this:
2528
2529 The regular expression compiler produces polymorphic opcodes. That
2530 is, the pattern adapts to the data and automatically switches to
2531 the Unicode character scheme when presented with Unicode data--or
2532 instead uses a traditional byte scheme when presented with byte
2533 data.
2534
2535 Sadly, a wide octal escape does not cause a switch, and in a string with
2536 no other multibyte characters, these octal escapes are treated as before.
2537 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2538 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2539 Unicode string.
2540
2541 I have not perpetrated such confusion in PCRE. Up till now, it took just
2542 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2543 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2544 translate to the appropriate multibyte character.
2545
2546 29. Applied some refactoring to reduce the number of warnings from Microsoft
2547 and Borland compilers. This has included removing the fudge introduced
2548 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2549 a warning about an unused variable.
2550
2551 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2552 characters since release 4.0, because Perl (from release 5.004) does not.
2553 [Or at least, is documented not to: some releases seem to be in conflict
2554 with the documentation.] However, when a pattern was studied with
2555 pcre_study() and all its branches started with \s, PCRE still included VT
2556 as a possible starting character. Of course, this did no harm; it just
2557 caused an unnecessary match attempt.
2558
2559 22. Removed a now-redundant internal flag bit that recorded the fact that case
2560 dependency changed within the pattern. This was once needed for "required
2561 byte" processing, but is no longer used. This recovers a now-scarce options
2562 bit. Also moved the least significant internal flag bit to the most-
2563 significant bit of the word, which was not previously used (hangover from
2564 the days when it was an int rather than a uint) to free up another bit for
2565 the future.
2566
2567 23. Added support for CRLF line endings as well as CR and LF. As well as the
2568 default being selectable at build time, it can now be changed at runtime
2569 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2570 specify that it is scanning data with non-default line endings.
2571
2572 24. Changed the definition of CXXLINK to make it agree with the definition of
2573 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2574
2575 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2576 recursions. This makes a big different to stack usage for some patterns.
2577
2578 26. If a subpattern containing a named recursion or subroutine reference such
2579 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2580 the space required for the compiled pattern went wrong and gave too small a
2581 value. Depending on the environment, this could lead to "Failed: internal
2582 error: code overflow at offset 49" or "glibc detected double free or
2583 corruption" errors.
2584
2585 27. Applied patches from Google (a) to support the new newline modes and (b) to
2586 advance over multibyte UTF-8 characters in GlobalReplace.
2587
2588 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2589 difference for some implementation of PCRE in some Windows version.
2590
2591 29. Added some extra testing facilities to pcretest:
2592
2593 \q<number> in a data line sets the "match limit" value
2594 \Q<number> in a data line sets the "match recursion limt" value
2595 -S <number> sets the stack size, where <number> is in megabytes
2596
2597 The -S option isn't available for Windows.
2598
2599
2600 Version 6.6 06-Feb-06
2601 ---------------------
2602
2603 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2604 in pcreposix.h. I have copied the definition from pcre.h.
2605
2606 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2607 because pcre.h is no longer a built file.
2608
2609 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2610 not normally included in the compiled code.
2611
2612
2613 Version 6.5 01-Feb-06
2614 ---------------------
2615
2616 1. When using the partial match feature with pcre_dfa_exec(), it was not
2617 anchoring the second and subsequent partial matches at the new starting
2618 point. This could lead to incorrect results. For example, with the pattern
2619 /1234/, partially matching against "123" and then "a4" gave a match.
2620
2621 2. Changes to pcregrep:
2622
2623 (a) All non-match returns from pcre_exec() were being treated as failures
2624 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2625 error message is output. Some extra information is given for the
2626 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2627 probably the only errors that are likely to be caused by users (by
2628 specifying a regex that has nested indefinite repeats, for instance).
2629 If there are more than 20 of these errors, pcregrep is abandoned.
2630
2631 (b) A binary zero was treated as data while matching, but terminated the
2632 output line if it was written out. This has been fixed: binary zeroes
2633 are now no different to any other data bytes.
2634
2635 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2636 used to set a locale for matching. The --locale=xxxx long option has
2637 been added (no short equivalent) to specify a locale explicitly on the
2638 pcregrep command, overriding the environment variables.
2639
2640 (d) When -B was used with -n, some line numbers in the output were one less
2641 than they should have been.
2642
2643 (e) Added the -o (--only-matching) option.
2644
2645 (f) If -A or -C was used with -c (count only), some lines of context were
2646 accidentally printed for the final match.
2647
2648 (g) Added the -H (--with-filename) option.
2649
2650 (h) The combination of options -rh failed to suppress file names for files
2651 that were found from directory arguments.
2652
2653 (i) Added the -D (--devices) and -d (--directories) options.
2654
2655 (j) Added the -F (--fixed-strings) option.
2656
2657 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2658
2659 (l) Added the --colo(u)r option.
2660
2661 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2662 is not present by default.
2663
2664 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2665 items such as (?R) or (?1), when the recursion could match a number of
2666 alternatives. If it matched one of the alternatives, but subsequently,
2667 outside the recursion, there was a failure, the code tried to back up into
2668 the recursion. However, because of the way PCRE is implemented, this is not
2669 possible, and the result was an incorrect result from the match.
2670
2671 In order to prevent this happening, the specification of recursion has
2672 been changed so that all such subpatterns are automatically treated as
2673 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2674
2675 4. I had overlooked the fact that, in some locales, there are characters for
2676 which isalpha() is true but neither isupper() nor islower() are true. In
2677 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2678 and ordfeminine) are like this. This affected the treatment of \w and \W
2679 when they appeared in character classes, but not when they appeared outside
2680 a character class. The bit map for "word" characters is now created
2681 separately from the results of isalnum() instead of just taking it from the
2682 upper, lower, and digit maps. (Plus the underscore character, of course.)
2683
2684 5. The above bug also affected the handling of POSIX character classes such as
2685 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2686 permanent tables. Instead, the bit maps for such a class were previously
2687 created as the appropriate unions of the upper, lower, and digit bitmaps.
2688 Now they are created by subtraction from the [[:word:]] class, which has
2689 its own bitmap.
2690
2691 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2692 It is created by subtracting the vertical space characters (\x09, \x0a,
2693 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2694 subtraction was done in the overall bitmap for a character class, meaning
2695 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2696 be recognized. This bug has been fixed.
2697
2698 7. Patches from the folks at Google:
2699
2700 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2701 real life, but is still worth protecting against".
2702
2703 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2704 regular expressions".
2705
2706 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2707 have it.
2708
2709 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2710 "configure" and the latter not, in order to fix a problem somebody had
2711 with compiling the Arg class on HP-UX.
2712
2713 (e) Improve the error-handling of the C++ wrapper a little bit.
2714
2715 (f) New tests for checking recursion limiting.
2716
2717 8. The pcre_memmove() function, which is used only if the environment does not
2718 have a standard memmove() function (and is therefore rarely compiled),
2719 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2720 returning a result (though PCRE never actually uses the result).
2721
2722 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2723 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2724 returned instead of calling malloc() with an overflowing number that would
2725 most likely cause subsequent chaos.
2726
2727 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2728
2729 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2730 with this option is matched, the nmatch and pmatch options of regexec() are
2731 ignored.
2732
2733 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2734 provided in case anyone wants to the the POSIX interface with UTF-8
2735 strings.
2736
2737 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2738 C++ linking (needed for some HP-UX environments).
2739
2740 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2741 (unused parameter) and in the pcre_printint() function (omitted "default"
2742 switch label when the default is to do nothing).
2743
2744 15. Added some code to make it possible, when PCRE is compiled as a C++
2745 library, to replace subject pointers for pcre_exec() with a smart pointer
2746 class, thus making it possible to process discontinuous strings.
2747
2748 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2749 much the same function. They were added by different people who were trying
2750 to make PCRE easy to compile on non-Unix systems. It has been suggested
2751 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2752 for compiling on Windows systems. I have therefore replaced it with
2753 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2754 defaults to "extern" for C or "extern C" for C++, which works fine on
2755 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2756 SCOPE with something explicit in config.h. In addition:
2757
2758 (a) pcreposix.h still had just "extern" instead of either of these macros;
2759 I have replaced it with PCRE_DATA_SCOPE.
2760
2761 (b) Functions such as _pcre_xclass(), which are internal to the library,
2762 but external in the C sense, all had PCRE_EXPORT in their definitions.
2763 This is apparently wrong for the Windows case, so I have removed it.
2764 (It makes no difference on Unix-like systems.)
2765
2766 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2767 of recursive calls to match(). This is different to MATCH_LIMIT because
2768 that limits the total number of calls to match(), not all of which increase
2769 the depth of recursion. Limiting the recursion depth limits the amount of
2770 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2771 when PCRE is compiled, and changed at run time. A patch from Google adds
2772 this functionality to the C++ interface.
2773
2774 18. Changes to the handling of Unicode character properties:
2775
2776 (a) Updated the table to Unicode 4.1.0.
2777
2778 (b) Recognize characters that are not in the table as "Cn" (undefined).
2779
2780 (c) I revised the way the table is implemented to a much improved format
2781 which includes recognition of ranges. It now supports the ranges that
2782 are defined in UnicodeData.txt, and it also amalgamates other
2783 characters into ranges. This has reduced the number of entries in the
2784 table from around 16,000 to around 3,000, thus reducing its size
2785 considerably. I realized I did not need to use a tree structure after
2786 all - a binary chop search is just as efficient. Having reduced the
2787 number of entries, I extended their size from 6 bytes to 8 bytes to
2788 allow for more data.
2789
2790 (d) Added support for Unicode script names via properties such as \p{Han}.
2791
2792 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2793 matching that character.
2794
2795 20. When matching a repeated Unicode property with a minimum greater than zero,
2796 (for example \pL{2,}), PCRE could look past the end of the subject if it
2797 reached it while seeking the minimum number of characters. This could
2798 happen only if some of the characters were more than one byte long, because
2799 there is a check for at least the minimum number of bytes.
2800
2801 21. Refactored the implementation of \p and \P so as to be more general, to
2802 allow for more different types of property in future. This has changed the
2803 compiled form incompatibly. Anybody with saved compiled patterns that use
2804 \p or \P will have to recompile them.
2805
2806 22. Added "Any" and "L&" to the supported property types.
2807
2808 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2809 but give a compile time error if the value is greater than 0xff.
2810
2811 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2812 accidentally not being installed or uninstalled.
2813
2814 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2815 made were to insert the current release number. This seemed silly, because
2816 it made things harder for people building PCRE on systems that don't run
2817 "configure". I have turned pcre.h into a distributed file, no longer built
2818 by "configure", with the version identification directly included. There is
2819 no longer a pcre.h.in file.
2820
2821 However, this change necessitated a change to the pcre-config script as
2822 well. It is built from pcre-config.in, and one of the substitutions was the
2823 release number. I have updated configure.ac so that ./configure now finds
2824 the release number by grepping pcre.h.
2825
2826 26. Added the ability to run the tests under valgrind.
2827
2828
2829 Version 6.4 05-Sep-05
2830 ---------------------
2831
2832 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2833 "--" to be printed when multiple files were scanned, even when none of the
2834 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2835 consider it to be a bug, and have restored the previous behaviour.
2836
2837 2. A couple of code tidies to get rid of compiler warnings.
2838
2839 3. The pcretest program used to cheat by referring to symbols in the library
2840 whose names begin with _pcre_. These are internal symbols that are not
2841 really supposed to be visible externally, and in some environments it is
2842 possible to suppress them. The cheating is now confined to including
2843 certain files from the library's source, which is a bit cleaner.
2844
2845 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2846 file's purpose clearer.
2847
2848 5. Reorganized pcre_ucp_findchar().
2849
2850
2851 Version 6.3 15-Aug-05
2852 ---------------------
2853
2854 1. The file libpcre.pc.in did not have general read permission in the tarball.
2855
2856 2. There were some problems when building without C++ support:
2857
2858 (a) If C++ support was not built, "make install" and "make test" still
2859 tried to test it.
2860
2861 (b) There were problems when the value of CXX was explicitly set. Some
2862 changes have been made to try to fix these, and ...
2863
2864 (c) --disable-cpp can now be used to explicitly disable C++ support.
2865
2866 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2867 backslash in a target when C++ was disabled. This confuses some
2868 versions of "make", apparently. Using an intermediate variable solves
2869 this. (Same for CPP_LOBJ.)
2870
2871 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2872 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2873 necessary on certain architectures.
2874
2875 4. Added a setting of -export-symbols-regex to the link command to remove
2876 those symbols that are exported in the C sense, but actually are local
2877 within the library, and not documented. Their names all begin with
2878 "_pcre_". This is not a perfect job, because (a) we have to except some
2879 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2880 available (and never for static libraries). I have made a note to try to
2881 find a way round (a) in the future.
2882
2883
2884 Version 6.2 01-Aug-05
2885 ---------------------
2886
2887 1. There was no test for integer overflow of quantifier values. A construction
2888 such as {1111111111111111} would give undefined results. What is worse, if
2889 a minimum quantifier for a parenthesized subpattern overflowed and became
2890 negative, the calculation of the memory size went wrong. This could have
2891 led to memory overwriting.
2892
2893 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2894
2895 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2896 operating environments where this matters.
2897
2898 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2899 PCRE options from within the C++ wrapper.
2900
2901 5. Named capturing subpatterns were not being correctly counted when a pattern
2902 was compiled. This caused two problems: (a) If there were more than 100
2903 such subpatterns, the calculation of the memory needed for the whole
2904 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2905 back references of the form \12, where the number was greater than 9, were
2906 not recognized as back references, even though there were sufficient
2907 previous subpatterns.
2908
2909 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2910 versions of gcc, e.g. 2.95.4.
2911
2912
2913 Version 6.1 21-Jun-05
2914 ---------------------
2915
2916 1. There was one reference to the variable "posix" in pcretest.c that was not
2917 surrounded by "#if !defined NOPOSIX".
2918
2919 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2920 the cross-check on the old pcre_info() function, for the benefit of the
2921 cut-down version of PCRE that is currently imported into Exim.
2922
2923 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2924 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2925 patterns that start (?i) but I don't think that matters. The use of (?i) is
2926 just an example; this all applies to the other options as well.
2927
2928 4. Since libtool seems to echo the compile commands it is issuing, the output
2929 from "make" can be reduced a bit by putting "@" in front of each libtool
2930 compile command.
2931
2932 5. Patch from the folks at Google for configure.in to be a bit more thorough
2933 in checking for a suitable C++ installation before trying to compile the
2934 C++ stuff. This should fix a reported problem when a compiler was present,
2935 but no suitable headers.
2936
2937 6. The man pages all had just "PCRE" as their title. I have changed them to
2938 be the relevant file name. I have also arranged that these names are
2939 retained in the file doc/pcre.txt, which is a concatenation in text format
2940 of all the man pages except the little individual ones for each function.
2941
2942 7. The NON-UNIX-USE file had not been updated for the different set of source
2943 files that come with release 6. I also added a few comments about the C++
2944 wrapper.
2945
2946
2947 Version 6.0 07-Jun-05
2948 ---------------------
2949
2950 1. Some minor internal re-organization to help with my DFA experiments.
2951
2952 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2953 didn't matter for the library itself when fully configured, but did matter
2954 when compiling without UCP support, or within Exim, where the ucp files are
2955 not imported.
2956
2957 3. Refactoring of the library code to split up the various functions into
2958 different source modules. The addition of the new DFA matching code (see
2959 below) to a single monolithic source would have made it really too
2960 unwieldy, quite apart from causing all the code to be include in a
2961 statically linked application, when only some functions are used. This is
2962 relevant even without the DFA addition now that patterns can be compiled in
2963 one application and matched in another.
2964
2965 The downside of splitting up is that there have to be some external
2966 functions and data tables that are used internally in different modules of
2967 the library but which are not part of the API. These have all had their
2968 names changed to start with "_pcre_" so that they are unlikely to clash
2969 with other external names.
2970
2971 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2972 a different (DFA) algorithm. Although it is slower than the original
2973 function, it does have some advantages for certain types of matching
2974 problem.
2975
2976 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2977 including restarting after a partial match.
2978
2979 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2980 defined when compiling for Windows was sent to me. I have put it into the
2981 code, though I have no means of testing or verifying it.
2982
2983 7. Added the pcre_refcount() auxiliary function.
2984
2985 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2986 match before or at the first newline in the subject string. In pcretest,
2987 the /f option on a pattern can be used to set this.
2988
2989 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2990 would behave wrongly. This has been present in PCRE since release 4.0.
2991
2992 10. A number of changes to the pcregrep command:
2993
2994 (a) Refactored how -x works; insert ^(...)$ instead of setting
2995 PCRE_ANCHORED and checking the length, in preparation for adding
2996 something similar for -w.
2997
2998 (b) Added the -w (match as a word) option.
2999
3000 (c) Refactored the way lines are read and buffered so as to have more
3001 than one at a time available.
3002
3003 (d) Implemented a pcregrep test script.
3004
3005 (e) Added the -M (multiline match) option. This allows patterns to match
3006 over several lines of the subject. The buffering ensures that at least
3007 8K, or the rest of the document (whichever is the shorter) is available
3008 for matching (and similarly the previous 8K for lookbehind assertions).
3009
3010 (f) Changed the --help output so that it now says
3011
3012 -w, --word-regex(p)
3013
3014 instead of two lines, one with "regex" and the other with "regexp"
3015 because that confused at least one person since the short forms are the
3016 same. (This required a bit of code, as the output is generated
3017 automatically from a table. It wasn't just a text change.)
3018
3019 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3020 option but starts with a hyphen. Could be a pattern or a path name
3021 starting with a hyphen, for instance.
3022
3023 (h) "-" can be given as a file name to represent stdin.
3024
3025 (i) When file names are being printed, "(standard input)" is used for
3026 the standard input, for compatibility with GNU grep. Previously
3027 "<stdin>" was used.
3028
3029 (j) The option --label=xxx can be used to supply a name to be used for
3030 stdin when file names are being printed. There is no short form.
3031
3032 (k) Re-factored the options decoding logic because we are going to add
3033 two more options that take data. Such options can now be given in four
3034 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3035
3036 (l) Added the -A, -B, and -C options for requesting that lines of context
3037 around matches be printed.
3038
3039 (m) Added the -L option to print the names of files that do not contain
3040 any matching lines, that is, the complement of -l.
3041
3042 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3043 continue to scan other files.
3044
3045 (o) The -s option was incorrectly implemented. For compatibility with other
3046 greps, it now suppresses the error message for a non-existent or non-
3047 accessible file (but not the return code). There is a new option called
3048 -q that suppresses the output of matching lines, which was what -s was
3049 previously doing.
3050
3051 (p) Added --include and --exclude options to specify files for inclusion
3052 and exclusion when recursing.
3053
3054 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3055 Hopefully, it now does.
3056
3057 12. Missing cast in pcre_study().
3058
3059 13. Added an "uninstall" target to the makefile.
3060
3061 14. Replaced "extern" in the function prototypes in Makefile.in with
3062 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3063 world, but is set differently for Windows.
3064
3065 15. Added a second compiling function called pcre_compile2(). The only
3066 difference is that it has an extra argument, which is a pointer to an
3067 integer error code. When there is a compile-time failure, this is set
3068 non-zero, in addition to the error test pointer being set to point to an
3069 error message. The new argument may be NULL if no error number is required
3070 (but then you may as well call pcre_compile(), which is now just a
3071 wrapper). This facility is provided because some applications need a
3072 numeric error indication, but it has also enabled me to tidy up the way
3073 compile-time errors are handled in the POSIX wrapper.
3074
3075 16. Added VPATH=.libs to the makefile; this should help when building with one
3076 prefix path and installing with another. (Or so I'm told by someone who
3077 knows more about this stuff than I do.)
3078
3079 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3080 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3081 match everything, including newlines. This is not POSIX-compatible, but
3082 somebody wanted the feature. From pcretest it can be activated by using
3083 both the P and the s flags.
3084
3085 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3086
3087 19. libpcre.pc was being incorrectly installed as executable.
3088
3089 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3090 it now also looks for '\r' so that it will work unmodified on Windows.
3091
3092 21. Added Google's contributed C++ wrapper to the distribution.
3093
3094 22. Added some untidy missing memory free() calls in pcretest, to keep
3095 Electric Fence happy when testing.
3096
3097
3098
3099 Version 5.0 13-Sep-04
3100 ---------------------
3101
3102 1. Internal change: literal characters are no longer packed up into items
3103 containing multiple characters in a single byte-string. Each character
3104 is now matched using a separate opcode. However, there may be more than one
3105 byte in the character in UTF-8 mode.
3106
3107 2. The pcre_callout_block structure has two new fields: pattern_position and
3108 next_item_length. These contain the offset in the pattern to the next match
3109 item, and its length, respectively.
3110
3111 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3112 insertion of callouts before each pattern item. Added the /C option to
3113 pcretest to make use of this.
3114
3115 4. On the advice of a Windows user, the lines
3116
3117 #if defined(_WIN32) || defined(WIN32)
3118 _setmode( _fileno( stdout ), 0x8000 );
3119 #endif /* defined(_WIN32) || defined(WIN32) */
3120
3121 have been added to the source of pcretest. This apparently does useful
3122 magic in relation to line terminators.
3123
3124 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3125 for the benefit of those environments where the "b" makes a difference.
3126
3127 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3128 to know about it. I have put a hack into configure.in that adds in code
3129 to set GCC=yes if CC=icc. This seems to end up at a point in the
3130 generated configure script that is early enough to affect the setting of
3131 compiler options, which is what is needed, but I have no means of testing
3132 whether it really works. (The user who reported this had patched the
3133 generated configure script, which of course I cannot do.)
3134
3135 LATER: After change 22 below (new libtool files), the configure script
3136 seems to know about icc (and also ecc). Therefore, I have commented out
3137 this hack in configure.in.
3138
3139 7. Added support for pkg-config (2 patches were sent in).
3140
3141 8. Negated POSIX character classes that used a combination of internal tables
3142 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3143 [[:^ascii]]. Typically, they would match almost any characters. The other
3144 POSIX classes were not broken in this way.
3145
3146 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3147 to find the match, as PCRE was deluded into thinking that the match had to
3148 start at the start point or following a newline. The same bug applied to
3149 patterns with negative forward assertions or any backward assertions
3150 preceding ".*" at the start, unless the pattern required a fixed first
3151 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3152
3153 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3154 starting at the last subject character, bytes beyond the end of the subject
3155 string were read.
3156
3157 11. Renamed the variable "class" as "classbits" to make life easier for C++
3158 users. (Previously there was a macro definition, but it apparently wasn't
3159 enough.)
3160
3161 12. Added the new field "tables" to the extra data so that tables can be passed
3162 in at exec time, or the internal tables can be re-selected. This allows
3163 a compiled regex to be saved and re-used at a later time by a different
3164 program that might have everything at different addresses.
3165
3166 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3167 -R library as well as a -L library.
3168
3169 14. The debugging options of pcretest (-d on the command line or D on a
3170 pattern) showed incorrect output for anything following an extended class
3171 that contained multibyte characters and which was followed by a quantifier.
3172
3173 15. Added optional support for general category Unicode character properties
3174 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3175 support. It adds about 90K to the size of the library. The meanings of the
3176 inbuilt class escapes such as \d and \s have NOT been changed.
3177
3178 16. Updated pcredemo.c to include calls to free() to release the memory for the
3179 compiled pattern.
3180
3181 17. The generated file chartables.c was being created in the source directory
3182 instead of in the building directory. This caused the build to fail if the
3183 source directory was different from the building directory, and was
3184 read-only.
3185
3186 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3187 file. No doubt somebody will tell me if they don't make sense... Also added
3188 Dan Mooney's comments about building on OpenVMS.
3189
3190 19. Added support for partial matching via the PCRE_PARTIAL option for
3191 pcre_exec() and the \P data escape in pcretest.
3192
3193 20. Extended pcretest with 3 new pattern features:
3194
3195 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3196 write the compiled pattern to the file whose name is "rest-of-line".
3197 This is a straight binary dump of the data, with the saved pointer to
3198 the character tables forced to be NULL. The study data, if any, is
3199 written too. After writing, pcretest reads a new pattern.
3200
3201 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3202 compiled pattern from the given file. There must not be any
3203 occurrences of "<" in the file name (pretty unlikely); if there are,
3204 pcretest will instead treat the initial "<" as a pattern delimiter.
3205 After reading in the pattern, pcretest goes on to read data lines as
3206 usual.
3207
3208 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3209 and 16-bit fields in a compiled pattern, to simulate a pattern that
3210 was compiled on a host of opposite endianness.
3211
3212 21. The pcre-exec() function can now cope with patterns that were compiled on
3213 hosts of opposite endianness, with this restriction:
3214
3215 As for any compiled expression that is saved and used later, the tables
3216 pointer field cannot be preserved; the extra_data field in the arguments
3217 to pcre_exec() should be used to pass in a tables address if a value
3218 other than the default internal tables were used at compile time.
3219
3220 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3221 now diagnosed as an error. Previously, most of the time, a negative number
3222 would have been treated as zero, but if in addition "ovector" was passed as
3223 NULL, a crash could occur.
3224
3225 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3226 new versions from the libtool 1.5 distribution (the last one is a copy of
3227 a file called libtool.m4). This seems to have fixed the need to patch
3228 "configure" to support Darwin 1.3 (which I used to do). However, I still
3229 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3230 workstation).
3231
3232 24. Changed the PCRE licence to be the more standard "BSD" licence.
3233
3234
3235 Version 4.5 01-Dec-03
3236 ---------------------
3237
3238 1. There has been some re-arrangement of the code for the match() function so
3239 that it can be compiled in a version that does not call itself recursively.
3240 Instead, it keeps those local variables that need separate instances for
3241 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3242 needs to "recurse". Keeping track of where control must go is done by means
3243 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3244 hide most of the details from the main code, and operates only if
3245 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3246 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3247 operating.
3248
3249 To make it easier for callers to provide specially tailored get/free
3250 functions for this usage, two new functions, pcre_stack_malloc, and
3251 pcre_stack_free, are used. They are always called in strict stacking order,
3252 and the size of block requested is always the same.
3253
3254 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3255 PCRE has been compiled to use the stack or the heap for recursion. The
3256 -C option of pcretest uses this to show which version is compiled.
3257
3258 A new data escape \S, is added to pcretest; it causes the amounts of store
3259 obtained and freed by both kinds of malloc/free at match time to be added
3260 to the output.
3261
3262 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3263 what's available on my current Linux desktop machine.
3264
3265 3. When matching a UTF-8 string, the test for a valid string at the start has
3266 been extended. If start_offset is not zero, PCRE now checks that it points
3267 to a byte that is the start of a UTF-8 character. If not, it returns
3268 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3269 this is necessary because there may be backward assertions in the pattern.
3270 When matching the same subject several times, it may save resources to use
3271 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3272
3273 4. The code for checking the validity of UTF-8 strings has been tightened so
3274 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3275 containing "overlong sequences".
3276
3277 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3278 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3279 should have been "&", but it just so happened that all the cases this let
3280 through by mistake were picked up later in the function.
3281
3282 6. I had used a variable called "isblank" - this is a C99 function, causing
3283 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3284
3285 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3286 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3287 so the version is shown; (c) stop "make test" echoing "./RunTest".
3288
3289 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3290
3291 9. The prototype for memmove() for systems that don't have it was using
3292 size_t, but the inclusion of the header that defines size_t was later. I've
3293 moved the #includes for the C headers earlier to avoid this.
3294
3295 10. Added some adjustments to the code to make it easier to compiler on certain
3296 special systems:
3297
3298 (a) Some "const" qualifiers were missing.
3299 (b) Added the macro EXPORT before all exported functions; by default this
3300 is defined to be empty.
3301 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3302 that it reads its output file name as an argument instead of writing
3303 to the standard output and assuming this can be redirected.
3304
3305 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3306 class containing characters with values greater than 255, PCRE compilation
3307 went into a loop.
3308
3309 12. A recursive reference to a subpattern that was within another subpattern
3310 that had a minimum quantifier of zero caused PCRE to crash. For example,
3311 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3312 recursion. If the recursively-called subpattern itself had a zero repeat,
3313 that was OK.
3314
3315 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3316 buffer into which it was copied (for escape processing) was still set at
3317 1024, so long lines caused crashes.
3318
3319 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3320 "internal error: code overflow...". This applied to any character class
3321 that was followed by a possessive quantifier.
3322
3323 15. Modified the Makefile to add libpcre.la as a prerequisite for
3324 libpcreposix.la because I was told this is needed for a parallel build to
3325 work.
3326
3327 16. If a pattern that contained .* following optional items at the start was
3328 studied, the wrong optimizing data was generated, leading to matching
3329 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3330 matching string must start with a or b or c. The correct conclusion for
3331 this pattern is that a match can start with any character.
3332
3333
3334 Version 4.4 13-Aug-03
3335 ---------------------
3336
3337 1. In UTF-8 mode, a character class containing characters with values between
3338 127 and 255 was not handled correctly if the compiled pattern was studied.
3339 In fixing this, I have also improved the studying algorithm for such
3340 classes (slightly).
3341
3342 2. Three internal functions had redundant arguments passed to them. Removal
3343 might give a very teeny performance improvement.
3344
3345 3. Documentation bug: the value of the capture_top field in a callout is *one
3346 more than* the number of the hightest numbered captured substring.
3347
3348 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3349 in incorrectly linking with a previously installed version. They now link
3350 explicitly with libpcre.la.
3351
3352 5. configure.in no longer needs to recognize Cygwin specially.
3353
3354 6. A problem in pcre.in for Windows platforms is fixed.
3355
3356 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3357 pcretest, it used to include the size of the study block as part of its
3358 output. Unfortunately, the structure contains a field that has a different
3359 size on different hardware architectures. This meant that the tests that
3360 showed this size failed. As the block is currently always of a fixed size,
3361 this information isn't actually particularly useful in pcretest output, so
3362 I have just removed it.
3363
3364 8. Three pre-processor statements accidentally did not start in column 1.
3365 Sadly, there are *still* compilers around that complain, even though
3366 standard C has not required this for well over a decade. Sigh.
3367
3368 9. In pcretest, the code for checking callouts passed small integers in the
3369 callout_data field, which is a void * field. However, some picky compilers
3370 complained about the casts involved for this on 64-bit systems. Now
3371 pcretest passes the address of the small integer instead, which should get
3372 rid of the warnings.
3373
3374 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3375 both compile and run time, and gives an error if an invalid UTF-8 sequence
3376 is found. There is a option for disabling this check in cases where the
3377 string is known to be correct and/or the maximum performance is wanted.
3378
3379 11. In response to a bug report, I changed one line in Makefile.in from
3380
3381 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3382 to
3383 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3384
3385 to look similar to other lines, but I have no way of telling whether this
3386 is the right thing to do, as I do not use Windows. No doubt I'll get told
3387 if it's wrong...
3388
3389
3390 Version 4.3 21-May-03
3391 ---------------------
3392
3393 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3394 Makefile.
3395
3396 2. Some refactoring to improve the quality of the code:
3397
3398 (i) The utf8_table... variables are now declared "const".
3399
3400 (ii) The code for \cx, which used the "case flipping" table to upper case
3401 lower case letters, now just substracts 32. This is ASCII-specific,
3402 but the whole concept of \cx is ASCII-specific, so it seems
3403 reasonable.
3404
3405 (iii) PCRE was using its character types table to recognize decimal and
3406 hexadecimal digits in the pattern. This is silly, because it handles
3407 only 0-9, a-f, and A-F, but the character types table is locale-
3408 specific, which means strange things might happen. A private
3409 table is now used for this - though it costs 256 bytes, a table is
3410 much faster than multiple explicit tests. Of course, the standard
3411 character types table is still used for matching digits in subject
3412 strings against \d.
3413
3414 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3415 ending in _t are). So I've renamed it as ESC_tee.
3416
3417 3. The first argument for regexec() in the POSIX wrapper should have been
3418 defined as "const".
3419
3420 4. Changed pcretest to use malloc() for its buffers so that they can be
3421 Electric Fenced for debugging.
3422
3423 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3424 to read one or more bytes before the start of the subject string. Often this
3425 had no effect on PCRE's behaviour, but in some circumstances it could
3426 provoke a segmentation fault.
3427
3428 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3429 to try to read one or more bytes before the start of the subject string.
3430
3431 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3432 UTF-8 support could misbehave in various ways if the subject string
3433 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3434 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3435 back over UTF-8 characters.)
3436
3437
3438 Version 4.2 14-Apr-03
3439 ---------------------
3440
3441 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3442
3443 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3444 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3445 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3446 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3447 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3448 and BUILD_EXEEXT
3449 Note: automatic setting of the BUILD variables is not yet working
3450 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3451 compile-time but not at link-time
3452 [LINK]: use for linking executables only
3453 make different versions for Windows and non-Windows
3454 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3455 libraries
3456 [LINK_FOR_BUILD]: new variable
3457 [OBJEXT]: use throughout
3458 [EXEEXT]: use throughout
3459 <winshared>: new target
3460 <wininstall>: new target
3461 <dftables.o>: use native compiler
3462 <dftables>: use native linker
3463 <install>: handle Windows platform correctly
3464 <clean>: ditto
3465 <check>: ditto
3466 copy DLL to top builddir before testing
3467
3468 As part of these changes, -no-undefined was removed again. This was reported
3469 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3470 in any case.
3471
3472 3. Some tidies to get rid of compiler warnings:
3473
3474 . In the match_data structure, match_limit was an unsigned long int, whereas
3475 match_call_count was an int. I've made them both unsigned long ints.
3476
3477 . In pcretest the fact that a const uschar * doesn't automatically cast to
3478 a void * provoked a warning.
3479
3480 . Turning on some more compiler warnings threw up some "shadow" variables
3481 and a few more missing casts.
3482
3483 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3484 option, a class that contained a single character with a value between 128
3485 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3486
3487 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3488 option, a class that contained several characters, but with at least one
3489 whose value was between 128 and 255 caused PCRE to crash.
3490
3491
3492 Version 4.1 12-Mar-03
3493 ---------------------
3494
3495 1. Compiling with gcc -pedantic found a couple of places where casts were
3496 needed, and a string in dftables.c that was longer than standard compilers are
3497 required to support.
3498
3499 2. Compiling with Sun's compiler found a few more places where the code could
3500 be tidied up in order to avoid warnings.
3501
3502 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3503 first of these names is deprecated in the latest Autoconf in favour of the name
3504 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3505 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3506 analogy I have changed it to CFLAGS_FOR_BUILD.
3507
3508 4. Added -no-undefined to the linking command in the Makefile, because this is
3509 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3510 linking step for the pcreposix library.
3511
3512 5. PCRE was failing to diagnose the case of two named groups with the same
3513 name.
3514
3515 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3516 literal character that is needed in the subject for a match, and scans along to
3517 ensure that it is present before embarking on the full matching process. This
3518 saves time in cases of nested unlimited repeats that are never going to match.
3519 Problem: the scan can take a lot of time if the subject is very long (e.g.
3520 megabytes), thus penalizing straightforward matches. It is now done only if the
3521 amount of subject to be scanned is less than 1000 bytes.
3522
3523 7. A lesser problem with the same optimization is that it was recording the
3524 first character of an anchored pattern as "needed", thus provoking a search
3525 right along the subject, even when the first match of the pattern was going to
3526 fail. The "needed" character is now not set for anchored patterns, unless it
3527 follows something in the pattern that is of non-fixed length. Thus, it still
3528 fulfils its original purpose of finding quick non-matches in cases of nested
3529 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3530
3531
3532 Version 4.0 17-Feb-03
3533 ---------------------
3534
3535 1. If a comment in an extended regex that started immediately after a meta-item
3536 extended to the end of string, PCRE compiled incorrect data. This could lead to
3537 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3538
3539 2. Moved to autoconf 2.53 and libtool 1.4.2.
3540
3541 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3542 the special perltest8 script is no longer needed - all the tests can be run
3543 from a single perltest script.
3544
3545 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3546 by \s. It has now been removed in PCRE. This means it isn't recognized as
3547 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3548 class [:space:] *does* include VT, thereby creating a mess.
3549
3550 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3551 space and tab.
3552
3553 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3554 its new features into the main test script, reducing the number of scripts.
3555
3556 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3557 were backward compatible, and made the (?i) apply to the whole pattern, as if
3558 /i were given. Now it behaves more logically, and applies the option setting
3559 only to what follows. PCRE has been changed to follow suit. However, if it
3560 finds options settings right at the start of the pattern, it extracts them into
3561 the global options, as before. Thus, they show up in the info data.
3562
3563 8. Added support for the \Q...\E escape sequence. Characters in between are
3564 treated as literals. This is slightly different from Perl in that $ and @ are
3565 also handled as literals inside the quotes. In Perl, they will cause variable
3566 interpolation. Note the following examples:
3567
3568 Pattern PCRE matches Perl matches
3569
3570 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3571 \Qabc\$xyz\E abc\$xyz abc\$xyz
3572 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3573
3574 For compatibility with Perl, \Q...\E sequences are recognized inside character
3575 classes as well as outside them.
3576
3577 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3578 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3579 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3580 signed/unsigned warnings.
3581
3582 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3583 option for pcretest, so I've replaced it by a simple function that does just
3584 that job.
3585
3586 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3587 "pcregrep -".
3588
3589 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3590 Java package. This provides some syntactic sugar for simple cases of what my
3591 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3592 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3593 item, you can use this simplified notation. Note that only makes sense with
3594 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3595 greediness, whatever the setting of the PCRE_UNGREEDY option.
3596
3597 13. A change of greediness default within a pattern was not taking effect at
3598 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3599 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3600 was abstracted outside.
3601
3602 14. PCRE now supports the \G assertion. It is true when the current matching
3603 position is at the start point of the match. This differs from \A when the
3604 starting offset is non-zero. Used with the /g option of pcretest (or similar
3605 code), it works in the same way as it does for Perl's /g option. If all
3606 alternatives of a regex begin with \G, the expression is anchored to the start
3607 match position, and the "anchored" flag is set in the compiled expression.
3608
3609 15. Some bugs concerning the handling of certain option changes within patterns
3610 have been fixed. These applied to options other than (?ims). For example,
3611 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3612 been the other way round. Some of this was related to change 7 above.
3613
3614 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3615 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3616 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3617 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3618
3619 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3620 mode. Unlike ".", it always matches newline, whatever the setting of
3621 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3622 assertions. Perl allows it, but it doesn't (in general) work because it can't
3623 calculate the length of the lookbehind. At least, that's the case for Perl
3624 5.8.0 - I've been told they are going to document that it doesn't work in
3625 future.
3626
3627 18. Added an error diagnosis for escapes that PCRE does not support: these are
3628 \L, \l, \N, \P, \p, \U, \u, and \X.
3629
3630 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3631 reading past the end of the pattern in cases such as /[abcd/.
3632
3633 20. PCRE was getting more memory than necessary for patterns with classes that
3634 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3635
3636 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3637 compiling PCRE for use with Virtual Pascal.
3638
3639 22. Small fix to the Makefile to make it work properly if the build is done
3640 outside the source tree.
3641
3642 23. Added a new extension: a condition to go with recursion. If a conditional
3643 subpattern starts with (?(R) the "true" branch is used if recursion has
3644 happened, whereas the "false" branch is used only at the top level.
3645
3646 24. When there was a very long string of literal characters (over 255 bytes
3647 without UTF support, over 250 bytes with UTF support), the computation of how
3648 much memory was required could be incorrect, leading to segfaults or other
3649 strange effects.
3650
3651 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3652 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3653 there was a subsequent back reference to those brackets. This meant that, for
3654 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3655 possible to check for precisely this case. All we can do is abandon the
3656 optimization if .* occurs inside capturing brackets when there are any back
3657 references whatsoever. (See below for a better fix that came later.)
3658
3659 26. The handling of the optimization for finding the first character of a
3660 non-anchored pattern, and for finding a character that is required later in the
3661 match were failing in some cases. This didn't break the matching; it just
3662 failed to optimize when it could. The way this is done has been re-implemented.
3663
3664 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3665
3666 28. Added a new feature that provides some of the functionality that Perl
3667 provides with (?{...}). The facility is termed a "callout". The way it is done
3668 in PCRE is for the caller to provide an optional function, by setting
3669 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3670 global variable. By default it is unset, which disables all calling out. To get
3671 the function called, the regex must include (?C) at appropriate points. This
3672 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3673 This provides a means of identifying different callout points. When PCRE
3674 reaches such a point in the regex, if pcre_callout has been set, the external
3675 function is called. It is provided with data in a structure called
3676 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3677 matching continues; if it returns a non-zero value, the match at the current
3678 point fails. However, backtracking will occur if possible. [This was changed
3679 later and other features added - see item 49 below.]
3680
3681 29. pcretest is upgraded to test the callout functionality. It provides a
3682 callout function that displays information. By default, it shows the start of
3683 the match and the current position in the text. There are some new data escapes
3684 to vary what happens:
3685
3686 \C+ in addition, show current contents of captured substrings
3687 \C- do not supply a callout function
3688 \C!n return 1 when callout number n is reached
3689 \C!n!m return 1 when callout number n is reached for the mth time
3690
3691 30. If pcregrep was called with the -l option and just a single file name, it
3692 output "<stdin>" if a match was found, instead of the file name.
3693
3694 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3695 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3696 pcre_exec(). This saves a malloc/free per call. The default value of
3697 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3698 when configuring.
3699
3700 32. The default maximum size of a compiled pattern is 64K. There have been a
3701 few cases of people hitting this limit. The code now uses macros to handle the
3702 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3703 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3704 configuring. Tests 2 and 5 work only with 2-byte links because they output
3705 debugging information about compiled patterns.
3706
3707 33. Internal code re-arrangements:
3708
3709 (a) Moved the debugging function for printing out a compiled regex into
3710 its own source file (printint.c) and used #include to pull it into
3711 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3712 separate copies.
3713
3714 (b) Defined the list of op-code names for debugging as a macro in
3715 internal.h so that it is next to the definition of the opcodes.
3716
3717 (c) Defined a table of op-code lengths for simpler skipping along compiled
3718 code. This is again a macro in internal.h so that it is next to the
3719 definition of the opcodes.
3720
3721 34. Added support for recursive calls to individual subpatterns, along the
3722 lines of Robin Houston's patch (but implemented somewhat differently).
3723
3724 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3725 allow it to read and process whole directories in Win32. This code was
3726 contributed by Lionel Fourquaux; it has not been tested by me.
3727
3728 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3729 used to name a group. Names consist of alphanumerics and underscores, and must
3730 be unique. Back references use the syntax (?P=name) and recursive calls use
3731 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3732 numbers. The function pcre_fullinfo() can be used after compilation to extract
3733 a name/number map. There are three relevant calls:
3734
3735 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3736 PCRE_INFO_NAMECOUNT yields the number of entries
3737 PCRE_INFO_NAMETABLE yields a pointer to the map.
3738
3739 The map is a vector of fixed-size entries. The size of each entry depends on
3740 the length of the longest name used. The first two bytes of each entry are the
3741 group number, most significant byte first. There follows the corresponding
3742 name, zero terminated. The names are in alphabetical order.
3743
3744 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3745 case instead of 255. Making it the same both with and without UTF-8 support
3746 means that the same test output works with both.
3747
3748 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3749 calling malloc() with a zero argument.
3750
3751 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3752 optimization. I've improved things by keeping a bitmap of backreferences with
3753 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3754 fact referenced, the optimization can be applied. It is unlikely that a
3755 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3756 the match to follow \n) will appear inside brackets with a number greater than
3757 31, but if it does, any back reference > 31 suppresses the optimization.
3758
3759 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3760 of disabling numbered capturing parentheses. Any opening parenthesis that is
3761 not followed by ? behaves as if it were followed by ?: but named parentheses
3762 can still be used for capturing (and they will acquire numbers in the usual
3763 way).
3764
3765 41. Redesigned the return codes from the match() function into yes/no/error so
3766 that errors can be passed back from deep inside the nested calls. A malloc
3767 failure while inside a recursive subpattern call now causes the
3768 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3769
3770 42. It is now possible to set a limit on the number of times the match()
3771 function is called in a call to pcre_exec(). This facility makes it possible to
3772 limit the amount of recursion and backtracking, though not in a directly
3773 obvious way, because the match() function is used in a number of different
3774 circumstances. The count starts from zero for each position in the subject
3775 string (for non-anchored patterns). The default limit is, for compatibility, a
3776 large number, namely 10 000 000. You can change this in two ways:
3777
3778 (a) When configuring PCRE before making, you can use --with-match-limit=n
3779 to set a default value for the compiled library.
3780
3781 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3782 a different value is set. See 45 below.
3783
3784 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3785
3786 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3787 of things that can be changed at compile time. The first argument specifies
3788 what is wanted and the second points to where the information is to be placed.
3789 The current list of available information is:
3790
3791 PCRE_CONFIG_UTF8
3792
3793 The output is an integer that is set to one if UTF-8 support is available;
3794 otherwise it is set to zero.
3795
3796 PCRE_CONFIG_NEWLINE
3797
3798 The output is an integer that it set to the value of the code that is used for
3799 newline. It is either LF (10) or CR (13).
3800
3801 PCRE_CONFIG_LINK_SIZE
3802
3803 The output is an integer that contains the number of bytes used for internal
3804 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3805
3806 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3807
3808 The output is an integer that contains the threshold above which the POSIX
3809 interface uses malloc() for output vectors. See item 31 above.
3810
3811 PCRE_CONFIG_MATCH_LIMIT
3812
3813 The output is an unsigned integer that contains the default limit of the number
3814 of match() calls in a pcre_exec() execution. See 42 above.
3815
3816 44. pcretest has been upgraded by the addition of the -C option. This causes it
3817 to extract all the available output from the new pcre_config() function, and to
3818 output it. The program then exits immediately.
3819
3820 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3821 order to support additional features. One way would have been to define
3822 pcre_exec2() (for example) with extra arguments, but this would not have been
3823 extensible, and would also have required all calls to the original function to
3824 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3825 is used for passing in "extra" data from pcre_study().
3826
3827 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3828 contains the following fields:
3829
3830 flags a bitmap indicating which of the following fields are set
3831 study_data opaque data from pcre_study()
3832 match_limit a way of specifying a limit on match() calls for a specific
3833 call to pcre_exec()
3834 callout_data data for callouts (see 49 below)
3835
3836 The flag bits are also defined in pcre.h, and are
3837
3838 PCRE_EXTRA_STUDY_DATA
3839 PCRE_EXTRA_MATCH_LIMIT
3840 PCRE_EXTRA_CALLOUT_DATA
3841
3842 The pcre_study() function now returns one of these new pcre_extra blocks, with
3843 the actual study data pointed to by the study_data field, and the
3844 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3845 before. That is, this change is entirely upwards-compatible and requires no
3846 change to existing code.
3847
3848 If you want to pass in additional data to pcre_exec(), you can either place it
3849 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3850 block.
3851
3852 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3853 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3854 times with different match limits, until it finds the minimum value needed for
3855 pcre_exec() to complete. The value is then output. This can be instructive; for
3856 most simple matches the number is quite small, but for pathological cases it
3857 gets very large very quickly.
3858
3859 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3860 returns the size of the data block pointed to by the study_data field in a
3861 pcre_extra block, that is, the value that was passed as the argument to
3862 pcre_malloc() when PCRE was getting memory in which to place the information
3863 created by pcre_study(). The fourth argument should point to a size_t variable.
3864 pcretest has been extended so that this information is shown after a successful
3865 pcre_study() call when information about the compiled regex is being displayed.
3866
3867 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3868 because what follows is always an absolute path. (Later: it turns out that this
3869 is more than cosmetic for MinGW, because it doesn't like empty path
3870 components.)
3871
3872 49. Some changes have been made to the callout feature (see 28 above):
3873
3874 (i) A callout function now has three choices for what it returns:
3875
3876 0 => success, carry on matching
3877 > 0 => failure at this point, but backtrack if possible
3878 < 0 => serious error, return this value from pcre_exec()
3879
3880 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3881 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3882 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3883 use by callout functions. It will never be used by PCRE itself.
3884
3885 (ii) The pcre_extra structure (see 45 above) has a void * field called
3886 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3887 pcre_callout_block structure has a field of the same name. The contents of
3888 the field passed in the pcre_extra structure are passed to the callout
3889 function in the corresponding field in the callout block. This makes it
3890 easier to use the same callout-containing regex from multiple threads. For
3891 testing, the pcretest program has a new data escape
3892
3893 \C*n pass the number n (may be negative) as callout_data
3894
3895 If the callout function in pcretest receives a non-zero value as
3896 callout_data, it returns that value.
3897
3898 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3899 there were some redundant $(CFLAGS) in commands that are now specified as
3900 $(LINK), which already includes $(CFLAGS).
3901
3902 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3903 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3904 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3905 one-byte characters throughout. Note that case-insensitive matching applies
3906 only to characters whose values are less than 256. PCRE doesn't support the
3907 notion of cases for higher-valued characters.
3908
3909 (i) A character class whose characters are all within 0-255 is handled as
3910 a bit map, and the map is inverted for negative classes. Previously, a
3911 character > 255 always failed to match such a class; however it should
3912 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3913
3914 (ii) A negated character class with a single character < 255 is coded as
3915 "not this character" (OP_NOT). This wasn't working properly when the test
3916 character was multibyte, either singly or repeated.
3917
3918 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3919 mode, for example: \x{100}{2,3}.
3920
3921 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3922 singly or repeated) now correctly test multibyte characters. However,
3923 PCRE doesn't recognize any characters with values greater than 255 as
3924 digits, spaces, or word characters. Such characters always match \D, \S,
3925 and \W, and never match \d, \s, or \w.
3926
3927 (v) Classes may now contain characters and character ranges with values
3928 greater than 255. For example: [ab\x{100}-\x{400}].
3929
3930 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3931 PCRE in UTF-8 mode.
3932
3933 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3934 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3935 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3936 value.)
3937
3938 53. The single man page has become too large. I have therefore split it up into
3939 a number of separate man pages. These also give rise to individual HTML pages;
3940 these are now put in a separate directory, and there is an index.html page that
3941 lists them all. Some hyperlinking between the pages has been installed.
3942
3943 54. Added convenience functions for handling named capturing parentheses.
3944
3945 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3946 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3947 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3948 are faulted.
3949
3950 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3951 calling configure. These values are used when compiling the dftables.c program
3952 which is run to generate the source of the default character tables. They
3953 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3954 you will need to set these values.
3955
3956 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3957
3958
3959 Version 3.9 02-Jan-02
3960 ---------------------
3961
3962 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3963
3964 2. If --disable-static was given, the building process failed when trying to
3965 build pcretest and pcregrep. (For some reason it was using libtool to compile
3966 them, which is not right, as they aren't part of the library.)
3967
3968
3969 Version 3.8 18-Dec-01
3970 ---------------------
3971
3972 1. The experimental UTF-8 code was completely screwed up. It was packing the
3973 bytes in the wrong order. How dumb can you get?
3974
3975
3976 Version 3.7 29-Oct-01
3977 ---------------------
3978
3979 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3980 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3981 this didn't happen under Solaris 8, where I normally test things.
3982
3983 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3984 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3985 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3986 this fix an hour or so after the initial 3.7 release.)
3987
3988
3989 Version 3.6 23-Oct-01
3990 ---------------------
3991
3992 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3993 offsets passed as NULL with zero offset count.
3994
3995 2. The config.guess and config.sub files had not been updated when I moved to
3996 the latest autoconf.
3997
3998
3999 Version 3.5 15-Aug-01
4000 ---------------------
4001
4002 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4003 had been forgotten.
4004
4005 2. By using declared but undefined structures, we can avoid using "void"
4006 definitions in pcre.h while keeping the internal definitions of the structures
4007 private.
4008
4009 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4010 user point of view, this means that both static and shared libraries are built
4011 by default, but this can be individually controlled. More of the work of
4012 handling this static/shared cases is now inside libtool instead of PCRE's make
4013 file.
4014
4015 4. The pcretest utility is now installed along with pcregrep because it is
4016 useful for users (to test regexs) and by doing this, it automatically gets
4017 relinked by libtool. The documentation has been turned into a man page, so
4018 there are now .1, .txt, and .html versions in /doc.
4019
4020 5. Upgrades to pcregrep:
4021 (i) Added long-form option names like gnu grep.
4022 (ii) Added --help to list all options with an explanatory phrase.
4023 (iii) Added -r, --recursive to recurse into sub-directories.
4024 (iv) Added -f, --file to read patterns from a file.
4025
4026 6. pcre_exec() was referring to its "code" argument before testing that
4027 argument for NULL (and giving an error if it was NULL).
4028
4029 7. Upgraded Makefile.in to allow for compiling in a different directory from
4030 the source directory.
4031
4032 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4033 options bits, the pointer it was passed was to an int instead of to an unsigned
4034 long int. This mattered only on 64-bit systems.
4035
4036 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4037 generated) instead of pcre.in, which it its source. Also made the same change
4038 in several of the .c files.
4039
4040 10. A new release of gcc defines printf() as a macro, which broke pcretest
4041 because it had an ifdef in the middle of a string argument for printf(). Fixed
4042 by using separate calls to printf().
4043
4044 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4045 script, to force use of CR or LF instead of \n in the source. On non-Unix
4046 systems, the value can be set in config.h.
4047
4048 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4049 absolute limit. Changed the text of the error message to make this clear, and
4050 likewise updated the man page.
4051
4052 13. The limit of 99 on the number of capturing subpatterns has been removed.
4053 The new limit is 65535, which I hope will not be a "real" limit.
4054
4055
4056 Version 3.4 22-Aug-00
4057 ---------------------
4058
4059 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4060
4061 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4062
4063
4064 Version 3.3 01-Aug-00
4065 ---------------------
4066
4067 1. If an octal character was given, but the value was greater than \377, it
4068 was not getting masked to the least significant bits, as documented. This could
4069 lead to crashes in some systems.
4070
4071 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4072 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4073
4074 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4075 These just pass their arguments on to (pcre_free)(), but they are provided
4076 because some uses of PCRE bind it to non-C systems that can call its functions,
4077 but cannot call free() or pcre_free() directly.
4078
4079 4. Add "make test" as a synonym for "make check". Corrected some comments in
4080 the Makefile.
4081
4082 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4083 Makefile.
4084
4085 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4086 command called pgrep for grepping around the active processes.
4087
4088 7. Added the beginnings of support for UTF-8 character strings.
4089
4090 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4091 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4092 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4093 out for the ar command.)
4094
4095
4096 Version 3.2 12-May-00
4097 ---------------------
4098
4099 This is purely a bug fixing release.
4100
4101 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4102 of ZA. This was just one example of several cases that could provoke this bug,
4103 which was introduced by change 9 of version 2.00. The code for breaking
4104 infinite loops after an iteration that matches an empty string was't working
4105 correctly.
4106
4107 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4108 when matched against abbab (for example). After matching an empty string, it
4109 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4110 caused it to match further down the string than it should.
4111
4112 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4113 was there because it doesn't seem to be needed, and it causes trouble on some
4114 systems, as it is not a Standard C header. It has been removed.
4115
4116 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4117 were reported on the Macintosh. The changes were from
4118
4119 while ((c = *(++ptr)) != 0 && c != '\n');
4120 to
4121 while ((c = *(++ptr)) != 0 && c != '\n') ;
4122
4123 Totally extraordinary, but if that's what it takes...
4124
4125 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4126 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4127 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4128 assumes the way PCRE uses memmove() (always moving upwards).
4129
4130 6. PCRE is being used in one environment where strchr() is not available. There
4131 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4132 faster code anyway.
4133
4134
4135 Version 3.1 09-Feb-00
4136 ---------------------
4137
4138 The only change in this release is the fixing of some bugs in Makefile.in for
4139 the "install" target:
4140
4141 (1) It was failing to install pcreposix.h.
4142
4143 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4144
4145
4146 Version 3.0 01-Feb-00
4147 ---------------------
4148
4149 1. Add support for the /+ modifier to perltest (to output $` like it does in
4150 pcretest).
4151
4152 2. Add support for the /g modifier to perltest.
4153
4154 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4155 matches null strings.
4156
4157 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4158 pattern. Perl treats empty patterns specially - it reuses the most recent
4159 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4160 effect.
4161
4162 5. The POSIX interface was broken in that it was just handing over the POSIX
4163 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4164 required a bigger vector, with some working space on the end. This means that
4165 the POSIX wrapper now has to get and free some memory, and copy the results.
4166
4167 6. Added some simple autoconf support, placing the test data and the
4168 documentation in separate directories, re-organizing some of the
4169 information files, and making it build pcre-config (a GNU standard). Also added
4170 libtool support for building PCRE as a shared library, which is now the
4171 default.
4172
4173 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4174 09 are not valid octal constants. Single digits will be used for minor values
4175 less than 10.
4176
4177 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4178 existing programs that set these in the POSIX interface can use PCRE without
4179 modification.
4180
4181 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4182 return all that pcre_info() returns, plus additional data. The pcre_info()
4183 function is retained for compatibility, but is considered to be obsolete.
4184
4185 10. Added experimental recursion feature (?R) to handle one common case that
4186 Perl 5.6 will be able to do with (?p{...}).
4187
4188 11. Added support for POSIX character classes like [:alpha:], which Perl is
4189 adopting.
4190
4191
4192 Version 2.08 31-Aug-99
4193 ----------------------
4194
4195 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4196 trying to match at the startoffset position, but instead was moving forward to
4197 the next newline as if a previous match had failed.
4198
4199 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4200 and could get into a loop if a null string was matched other than at the start
4201 of the subject.
4202
4203 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4204 be distinguished at compile time, and for completeness also added PCRE_DATE.
4205
4206 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4207 in GnuWin32 environments.
4208
4209
4210 Version 2.07 29-Jul-99
4211 ----------------------
4212
4213 1. The documentation is now supplied in plain text form and HTML as well as in
4214 the form of man page sources.
4215
4216 2. C++ compilers don't like assigning (void *) values to other pointer types.
4217 In particular this affects malloc(). Although there is no problem in Standard
4218 C, I've put in casts to keep C++ compilers happy.
4219
4220 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4221 should be (const char *).
4222
4223 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4224 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4225 However, I haven't made this a standard facility. The documentation doesn't
4226 mention it, and the Makefile doesn't support it.
4227
4228 5. The Makefile now contains an "install" target, with editable destinations at
4229 the top of the file. The pcretest program is not installed.
4230
4231 6. pgrep -V now gives the PCRE version number and date.
4232
4233 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4234 causing the entire string to be ignored, instead of just the last character.
4235
4236 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4237 non-matching string, it can take a very, very long time, even for strings of
4238 quite modest length, because of the nested recursion. PCRE now does better in
4239 some of these cases. It does this by remembering the last required literal
4240 character in the pattern, and pre-searching the subject to ensure it is present
4241 before running the real match. In other words, it applies a heuristic to detect
4242 some types of certain failure quickly, and in the above example, if presented
4243 with a string that has no trailing " it gives "no match" very quickly.
4244
4245 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4246 other alternatives are tried instead.
4247
4248
4249 Version 2.06 09-Jun-99
4250 ----------------------
4251
4252 1. Change pcretest's output for amount of store used to show just the code
4253 space, because the remainder (the data block) varies in size between 32-bit and
4254 64-bit systems.
4255
4256 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4257 start matching at. This allows lookbehinds to work when searching for multiple
4258 occurrences in a string.
4259
4260 3. Added additional options to pcretest for testing multiple occurrences:
4261
4262 /+ outputs the rest of the string that follows a match
4263 /g loops for multiple occurrences, using the new startoffset argument
4264 /G loops for multiple occurrences by passing an incremented pointer
4265
4266 4. PCRE wasn't doing the "first character" optimization for patterns starting
4267 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4268 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4269 the letter 'x'. On long subject strings, this gives a significant speed-up.
4270
4271
4272 Version 2.05 21-Apr-99
4273 ----------------------
4274
4275 1. Changed the type of magic_number from int to long int so that it works
4276 properly on 16-bit systems.
4277
4278 2. Fixed a bug which caused patterns starting with .* not to work correctly
4279 when the subject string contained newline characters. PCRE was assuming
4280 anchoring for such patterns in all cases, which is not correct because .* will
4281 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4282 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4283 must be retried after every newline in the subject.
4284
4285
4286 Version 2.04 18-Feb-99
4287 ----------------------
4288
4289 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4290 computation of the store needed to hold the pattern was incorrect (too large).
4291 If such patterns were nested a few deep, this could multiply and become a real
4292 problem.
4293
4294 2. Added /M option to pcretest to show the memory requirement of a specific
4295 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4296
4297 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4298 compiled in such a way that the backtracking after subsequent failure was
4299 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4300 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4301
4302
4303 Version 2.03 02-Feb-99
4304 ----------------------
4305
4306 1. Fixed typo and small mistake in man page.
4307
4308 2. Added 4th condition (GPL supersedes if conflict) and created separate
4309 LICENCE file containing the conditions.
4310
4311 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4312 Perl, that is the internal \ allows the delimiter to be included in the
4313 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4314 the final delimiter, add \ to the end of the pattern (to test the error).
4315
4316 4. Added the convenience functions for extracting substrings after a successful
4317 match. Updated pcretest to make it able to test these functions.
4318
4319
4320 Version 2.02 14-Jan-99
4321 ----------------------
4322
4323 1. Initialized the working variables associated with each extraction so that
4324 their saving and restoring doesn't refer to uninitialized store.
4325
4326 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4327 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4328 fix the problem.
4329
4330 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4331 calls, and wasn't printing the correct value for compiling calls. Increased the
4332 default value of LOOPREPEAT, and the number of significant figures in the
4333 times.
4334
4335 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4336
4337 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4338 a building problem on Windows NT with a FAT file system.
4339
4340
4341 Version 2.01 21-Oct-98
4342 ----------------------
4343
4344 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4345 to character tables built by pcre_maketables() in the current locale. If NULL
4346 is passed, the default tables are used.
4347
4348
4349 Version 2.00 24-Sep-98
4350 ----------------------
4351
4352 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4353 it any more.
4354
4355 2. Allow quantification of (?>) groups, and make it work correctly.
4356
4357 3. The first character computation wasn't working for (?>) groups.
4358
4359 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4360 end of the subject) and add 5.005's \z, which really does match only at the
4361 very end of the subject.
4362
4363 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4364
4365 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4366 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4367 localized options. All options to pcre_study() were also removed.
4368
4369 7. Add other new features from 5.005:
4370
4371 $(?<= positive lookbehind
4372 $(?<! negative lookbehind
4373 (?imsx-imsx) added the unsetting capability
4374 such a setting is global if at outer level; local otherwise
4375 (?imsx-imsx:) non-capturing groups with option setting
4376 (?(cond)re|re) conditional pattern matching
4377
4378 A backreference to itself in a repeated group matches the previous
4379 captured string.
4380
4381 8. General tidying up of studying (both automatic and via "study")
4382 consequential on the addition of new assertions.
4383
4384 9. As in 5.005, unlimited repeated groups that could match an empty substring
4385 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4386 runtime if any iteration does actually match an empty substring.
4387
4388 10. Include the RunTest script in the distribution.
4389
4390 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4391 discrepancies, some of which were old and were also with respect to 5.004. They
4392 have now been fixed.
4393
4394
4395 Version 1.09 28-Apr-98
4396 ----------------------
4397
4398 1. A negated single character class followed by a quantifier with a minimum
4399 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4400 program crashes, or just wrong answers. This did not apply to negated classes
4401 containing more than one character, or to minima other than one.
4402
4403
4404 Version 1.08 27-Mar-98
4405 ----------------------
4406
4407 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4408
4409 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4410 latter must appear before anything that relies on it in the pattern.
4411
4412
4413 Version 1.07 16-Feb-98
4414 ----------------------
4415
4416 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4417 repeat of a potentially empty string).
4418
4419
4420 Version 1.06 23-Jan-98
4421 ----------------------
4422
4423 1. Added Markus Oberhumer's little patches for C++.
4424
4425 2. Literal strings longer than 255 characters were broken.
4426
4427
4428 Version 1.05 23-Dec-97
4429 ----------------------
4430
4431 1. Negated character classes containing more than one character were failing if
4432 PCRE_CASELESS was set at run time.
4433
4434
4435 Version 1.04 19-Dec-97
4436 ----------------------
4437
4438 1. Corrected the man page, where some "const" qualifiers had been omitted.
4439
4440 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4441 input syntax.
4442
4443 3. Fixed memory leak which occurred when a regex with back references was
4444 matched with an offsets vector that wasn't big enough. The temporary memory
4445 that is used in this case wasn't being freed if the match failed.
4446
4447 4. Tidied pcretest to ensure it frees memory that it gets.
4448
4449 5. Temporary memory was being obtained in the case where the passed offsets
4450 vector was exactly big enough.
4451
4452 6. Corrected definition of offsetof() from change 5 below.
4453
4454 7. I had screwed up change 6 below and broken the rules for the use of
4455 setjmp(). Now fixed.
4456
4457
4458 Version 1.03 18-Dec-97
4459 ----------------------
4460
4461 1. A erroneous regex with a missing opening parenthesis was correctly
4462 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4463 on some systems.
4464
4465 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4466 it was reported that one broken compiler failed on the former because "code" is
4467 also an independent variable.
4468
4469 3. The erroneous regex a[]b caused an array overrun reference.
4470
4471 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4472 fail on data ending with that character. (It was going on too far, and checking
4473 the next character, typically a binary zero.) This was specific to the
4474 optimized code for single-character negative classes.
4475
4476 5. Added a contributed patch from the TIN world which does the following:
4477
4478 + Add an undef for memmove, in case the the system defines a macro for it.
4479
4480 + Add a definition of offsetof(), in case there isn't one. (I don't know
4481 the reason behind this - offsetof() is part of the ANSI standard - but
4482 it does no harm).
4483
4484 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4485 most of the places where whitespace preceded '#'. I have given up and
4486 allowed the remaining 2 cases to be at the margin.
4487
4488 + Rename some variables in pcre to eliminate shadowing. This seems very
4489 pedantic, but does no harm, of course.
4490
4491 6. Moved the call to setjmp() into its own function, to get rid of warnings
4492 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4493
4494 7. Constructs such as \d{8,} were compiling into the equivalent of
4495 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4496 outcome, but in this particular case used more store than had been allocated,
4497 which caused the bug to be discovered because it threw up an internal error.
4498
4499 8. The debugging code in both pcre and pcretest for outputting the compiled
4500 form of a regex was going wrong in the case of back references followed by
4501 curly-bracketed repeats.
4502
4503
4504 Version 1.02 12-Dec-97
4505 ----------------------
4506
4507 1. Typos in pcre.3 and comments in the source fixed.
4508
4509 2. Applied a contributed patch to get rid of places where it used to remove
4510 'const' from variables, and fixed some signed/unsigned and uninitialized
4511 variable warnings.
4512
4513 3. Added the "runtest" target to Makefile.
4514
4515 4. Set default compiler flag to -O2 rather than just -O.
4516
4517
4518 Version 1.01 19-Nov-97
4519 ----------------------
4520
4521 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4522 like /([ab]*)*/, that is, for classes with more than one character in them.
4523
4524 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4525 as /((?>a*))*/ (a PCRE_EXTRA facility).
4526
4527
4528 Version 1.00 18-Nov-97
4529 ----------------------
4530
4531 1. Added compile-time macros to support systems such as SunOS4 which don't have
4532 memmove() or strerror() but have other things that can be used instead.
4533
4534 2. Arranged that "make clean" removes the executables.
4535
4536
4537 Version 0.99 27-Oct-97
4538 ----------------------
4539
4540 1. Fixed bug in code for optimizing classes with only one character. It was
4541 initializing a 32-byte map regardless, which could cause it to run off the end
4542 of the memory it had got.
4543
4544 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4545
4546
4547 Version 0.98 22-Oct-97
4548 ----------------------
4549
4550 1. Fixed bug in code for handling temporary memory usage when there are more
4551 back references than supplied space in the ovector. This could cause segfaults.
4552
4553
4554 Version 0.97 21-Oct-97
4555 ----------------------
4556
4557 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4558
4559 2. Optimized negated single characters not to use a bit map.
4560
4561 3. Brought error texts together as macro definitions; clarified some of them;
4562 fixed one that was wrong - it said "range out of order" when it meant "invalid
4563 escape sequence".
4564
4565 4. Changed some char * arguments to const char *.
4566
4567 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4568
4569 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4570 pcretest.
4571
4572
4573 Version 0.96 16-Oct-97
4574 ----------------------
4575
4576 1. Added a simple "pgrep" utility to the distribution.
4577
4578 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4579 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4580 where "ddd" means "one or more decimal digits".
4581
4582 3. Fixed serious bug. If a pattern had a back reference, but the call to
4583 pcre_exec() didn't supply a large enough ovector to record the related
4584 identifying subpattern, the match always failed. PCRE now remembers the number
4585 of the largest back reference, and gets some temporary memory in which to save
4586 the offsets during matching if necessary, in order to ensure that
4587 backreferences always work.
4588
4589 4. Increased the compatibility with Perl in a number of ways:
4590
4591 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4592 to request this handling. The option can be set at compile or exec time.
4593
4594 (b) $ matches before a terminating newline by default; an option
4595 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4596 mode). The option can be set at compile or exec time.
4597
4598 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4599 the same as Perl's. If the decimal number it represents is less than 10
4600 or there aren't that many previous left capturing parentheses, an octal
4601 escape is read. Inside a character class, it's always an octal escape,
4602 even if it is a single digit.
4603
4604 (d) An escaped but undefined alphabetic character is taken as a literal,
4605 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4606 escapes.
4607
4608 (e) {0} is now permitted. (The previous item is removed from the compiled
4609 pattern).
4610
4611 5. Changed all the names of code files so that the basic parts are no longer
4612 than 10 characters, and abolished the teeny "globals.c" file.
4613
4614 6. Changed the handling of character classes; they are now done with a 32-byte
4615 bit map always.
4616
4617 7. Added the -d and /D options to pcretest to make it possible to look at the
4618 internals of compilation without having to recompile pcre.
4619
4620
4621 Version 0.95 23-Sep-97
4622 ----------------------
4623
4624 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4625 \x20 at the start of a run of normal characters. These were being treated as
4626 real characters, instead of the source characters being re-checked.
4627
4628
4629 Version 0.94 18-Sep-97
4630 ----------------------
4631
4632 1. The functions are now thread-safe, with the caveat that the global variables
4633 containing pointers to malloc() and free() or alternative functions are the
4634 same for all threads.
4635
4636 2. Get pcre_study() to generate a bitmap of initial characters for non-
4637 anchored patterns when this is possible, and use it if passed to pcre_exec().
4638
4639
4640 Version 0.93 15-Sep-97
4641 ----------------------
4642
4643 1. /(b)|(:+)/ was computing an incorrect first character.
4644
4645 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4646 but not actually doing anything yet.
4647
4648 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4649 as Perl does (e.g. [-az] or [az-]).
4650
4651 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4652 all possible positions.
4653
4654 5. Split up into different modules to avoid including unneeded functions in a
4655 compiled binary. However, compile and exec are still in one module. The "study"
4656 function is split off.
4657
4658 6. The character tables are now in a separate module whose source is generated
4659 by an auxiliary program - but can then be edited by hand if required. There are
4660 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4661 toupper() in the code.
4662
4663 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4664 make them global. Abolish the function for setting them, as the caller can now
4665 set them directly.
4666
4667
4668 Version 0.92 11-Sep-97
4669 ----------------------
4670
4671 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4672 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4673
4674 2. Caseless matching was not working in character classes if the characters in
4675 the pattern were in upper case.
4676
4677 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4678
4679 4. Make PCRE_ANCHORED public and accept as a compile option.
4680
4681 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4682 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4683 pass them.
4684
4685 6. Give an error if bad option bits passed at compile or run time.
4686
4687 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4688 pcretest to cause it to pass that flag.
4689
4690 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4691 options, and the first character, if set.
4692
4693 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4694
4695
4696 Version 0.91 10-Sep-97
4697 ----------------------
4698
4699 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4700 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4701
4702 2. PCRE was looping on encountering an indefinitely repeated back reference to
4703 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4704 Perl does - treats the match as successful.
4705
4706 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5