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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 922 - (show annotations) (download)
Mon Feb 20 18:44:42 2012 UTC (2 years, 1 month ago) by ph10
File size: 227503 byte(s)
Set PCRE_EXTRA_USED_JIT when JIT was actually used at runtime. Add /S++ and
-s++ to pcretest to show whether JIT was used or not. 

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

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12