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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 919 - (show annotations)
Fri Feb 17 11:48:02 2012 UTC (3 years, 5 months ago) by ph10
File size: 227161 byte(s)
Error occurred while calculating annotation data.
Fix partial match issues with CRLF and \r, and update tests.
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
37 Version 8.30 04-February-2012
38 -----------------------------
39
40 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
41 name is defined in ctype.h.
42
43 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
44 only in quite long subpatterns.
45
46 3. Removed the function pcre_info(), which has been obsolete and deprecated
47 since it was replaced by pcre_fullinfo() in February 2000.
48
49 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
50 match a (*MARK), and the match failed at the start of the subject, a
51 reference to memory before the start of the subject could occur. This bug
52 was introduced by fix 17 of release 8.21.
53
54 5. A reference to an unset group with zero minimum repetition was giving
55 totally wrong answers (in non-JavaScript-compatibility mode). For example,
56 /(another)?(\1?)test/ matched against "hello world test". This bug was
57 introduced in release 8.13.
58
59 6. Add support for 16-bit character strings (a large amount of work involving
60 many changes and refactorings).
61
62 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
63 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
64 from a file.
65
66 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
67 rounding is not applied in this particular case).
68
69 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
70 if they appear, or are escaped, in patterns.
71
72 10. Get rid of a number of -Wunused-but-set-variable warnings.
73
74 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
75 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
76 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
77 also returns the mark "x". This bug applied to capturing parentheses,
78 non-capturing parentheses, and atomic parentheses. It also applied to some
79 assertions.
80
81 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
82 information out of configure.ac instead of relying on pcre.h.generic, which
83 is not stored in the repository.
84
85 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
86 -lreadline.
87
88 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
89
90 15. Applied Graycode's patch to put the top-level frame on the stack rather
91 than the heap when not using the stack for recursion. This gives a
92 performance improvement in many cases when recursion is not deep.
93
94 16. Experimental code added to "pcretest -C" to output the stack frame size.
95
96
97 Version 8.21 12-Dec-2011
98 ------------------------
99
100 1. Updating the JIT compiler.
101
102 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
103 are added as well.
104
105 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
106 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
107 calling _pcre_jit_exec. Some extra comments are added.
108
109 4. (*MARK) settings inside atomic groups that do not contain any capturing
110 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
111 was introduced by change 18 for 8.20.
112
113 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
114 ECMA-262 standard.
115
116 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
117 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
118 This bug was probably introduced by change 9 of 8.13.
119
120 7. While fixing 6 above, I noticed that a number of other items were being
121 incorrectly rejected as "not fixed length". This arose partly because newer
122 opcodes had not been added to the fixed-length checking code. I have (a)
123 corrected the bug and added tests for these items, and (b) arranged for an
124 error to occur if an unknown opcode is encountered while checking for fixed
125 length instead of just assuming "not fixed length". The items that were
126 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
127 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
128 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
129
130 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
131 being incorrectly compiled and would have given unpredicatble results.
132
133 9. A possessively repeated subpattern with minimum repeat count greater than
134 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
135 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
136 the first (A) could occur when it should not.
137
138 10. Add a cast and remove a redundant test from the code.
139
140 11. JIT should use pcre_malloc/pcre_free for allocation.
141
142 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
143 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
144 is anything other than /usr, -L is still shown).
145
146 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
147
148 14. Perl does not support \N without a following name in a [] class; PCRE now
149 also gives an error.
150
151 15. If a forward reference was repeated with an upper limit of around 2000,
152 it caused the error "internal error: overran compiling workspace". The
153 maximum number of forward references (including repeats) was limited by the
154 internal workspace, and dependent on the LINK_SIZE. The code has been
155 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
156 the default depends on LINK_SIZE. There is a new upper limit (for safety)
157 of around 200,000 forward references. While doing this, I also speeded up
158 the filling in of repeated forward references.
159
160 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
161 incorrectly expecting the subject to contain another "a" after the start.
162
163 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
164 in the match, the SKIP should be ignored. This was not happening; instead
165 the SKIP was being treated as NOMATCH. For patterns such as
166 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
167 tested.
168
169 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
170 now much more compatible with Perl, in particular in cases where the result
171 is a non-match for a non-anchored pattern. For example, if
172 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
173 "m", where previously it did not return a name. A side effect of this
174 change is that for partial matches, the last encountered mark name is
175 returned, as for non matches. A number of tests that were previously not
176 Perl-compatible have been moved into the Perl-compatible test files. The
177 refactoring has had the pleasing side effect of removing one argument from
178 the match() function, thus reducing its stack requirements.
179
180 19. If the /S+ option was used in pcretest to study a pattern using JIT,
181 subsequent uses of /S (without +) incorrectly behaved like /S+.
182
183 21. Retrieve executable code size support for the JIT compiler and fixing
184 some warnings.
185
186 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
187 not work when the shorter character appeared right at the end of the
188 subject string.
189
190 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
191 systems.
192
193 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
194 output it when the /M option is used in pcretest.
195
196 25. The CheckMan script was not being included in the distribution. Also, added
197 an explicit "perl" to run Perl scripts from the PrepareRelease script
198 because this is reportedly needed in Windows.
199
200 26. If study data was being save in a file and studying had not found a set of
201 "starts with" bytes for the pattern, the data written to the file (though
202 never used) was taken from uninitialized memory and so caused valgrind to
203 complain.
204
205 27. Updated RunTest.bat as provided by Sheri Pierce.
206
207 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
208
209 29. Computation of memory usage for the table of capturing group names was
210 giving an unnecessarily large value.
211
212
213 Version 8.20 21-Oct-2011
214 ------------------------
215
216 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
217 a POSIX class. After further experiments with Perl, which convinced me that
218 Perl has bugs and confusions, a closing square bracket is no longer allowed
219 in a POSIX name. This bug also affected patterns with classes that started
220 with full stops.
221
222 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
223 captured substring, but while checking the failing first alternative,
224 substring 1 is temporarily captured. If the output vector supplied to
225 pcre_exec() was not big enough for this capture, the yield of the function
226 was still zero ("insufficient space for captured substrings"). This cannot
227 be totally fixed without adding another stack variable, which seems a lot
228 of expense for a edge case. However, I have improved the situation in cases
229 such as /(a)(b)x|abc/ matched against "abc", where the return code
230 indicates that fewer than the maximum number of slots in the ovector have
231 been set.
232
233 3. Related to (2) above: when there are more back references in a pattern than
234 slots in the output vector, pcre_exec() uses temporary memory during
235 matching, and copies in the captures as far as possible afterwards. It was
236 using the entire output vector, but this conflicts with the specification
237 that only 2/3 is used for passing back captured substrings. Now it uses
238 only the first 2/3, for compatibility. This is, of course, another edge
239 case.
240
241 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
242 main code base, and can be used by building with --enable-jit. When this is
243 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
244 runtime --no-jit option is given.
245
246 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
247 ovector, the return from the function was zero, implying that there were
248 other matches that did not fit. The correct "exactly full" value is now
249 returned.
250
251 6. If a subpattern that was called recursively or as a subroutine contained
252 (*PRUNE) or any other control that caused it to give a non-standard return,
253 invalid errors such as "Error -26 (nested recursion at the same subject
254 position)" or even infinite loops could occur.
255
256 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
257 computing the minimum length on reaching *ACCEPT, and so ended up with the
258 wrong value of 1 rather than 0. Further investigation indicates that
259 computing a minimum subject length in the presence of *ACCEPT is difficult
260 (think back references, subroutine calls), and so I have changed the code
261 so that no minimum is registered for a pattern that contains *ACCEPT.
262
263 8. If (*THEN) was present in the first (true) branch of a conditional group,
264 it was not handled as intended. [But see 16 below.]
265
266 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
267 Sheri Pierce.
268
269 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
270 the first byte in a match must be "a".
271
272 11. Change 17 for 8.13 increased the recursion depth for patterns like
273 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
274 pattern contains any instances of (*THEN). If it does not, the old
275 optimizations are restored. It would be nice to do this on a per-group
276 basis, but at the moment that is not feasible.
277
278 12. In some environments, the output of pcretest -C is CRLF terminated. This
279 broke RunTest's code that checks for the link size. A single white space
280 character after the value is now allowed for.
281
282 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
283 For "fr", it uses the Windows-specific input and output files.
284
285 14. If (*THEN) appeared in a group that was called recursively or as a
286 subroutine, it did not work as intended. [But see next item.]
287
288 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
289 pattern fragments (but not containing any | characters). If A and B are
290 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
291 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
292 D. In other words, Perl considers parentheses that do not contain any |
293 characters to be part of a surrounding alternative, whereas PCRE was
294 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
295 differently. PCRE now behaves in the same way as Perl, except in the case
296 of subroutine/recursion calls such as (?1) which have in any case always
297 been different (but PCRE had them first :-).
298
299 16. Related to 15 above: Perl does not treat the | in a conditional group as
300 creating alternatives. Such a group is treated in the same way as an
301 ordinary group without any | characters when processing (*THEN). PCRE has
302 been changed to match Perl's behaviour.
303
304 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
305 RunGrepTest script failed.
306
307 18. Change 22 for version 13 caused atomic groups to use more stack. This is
308 inevitable for groups that contain captures, but it can lead to a lot of
309 stack use in large patterns. The old behaviour has been restored for atomic
310 groups that do not contain any capturing parentheses.
311
312 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
313 suppress the check for a minimum subject length at run time. (If it was
314 given to pcre_exec() or pcre_dfa_exec() it did work.)
315
316 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
317 fail to work when decoding hex characters in data strings in EBCDIC
318 environments.
319
320 21. It appears that in at least one Mac OS environment, the isxdigit() function
321 is implemented as a macro that evaluates to its argument more than once,
322 contravening the C 90 Standard (I haven't checked a later standard). There
323 was an instance in pcretest which caused it to go wrong when processing
324 \x{...} escapes in subject strings. The has been rewritten to avoid using
325 things like p++ in the argument of isxdigit().
326
327
328 Version 8.13 16-Aug-2011
329 ------------------------
330
331 1. The Unicode data tables have been updated to Unicode 6.0.0.
332
333 2. Two minor typos in pcre_internal.h have been fixed.
334
335 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
336 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
337 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
338
339 4. There were a number of related bugs in the code for matching backrefences
340 caselessly in UTF-8 mode when codes for the characters concerned were
341 different numbers of bytes. For example, U+023A and U+2C65 are an upper
342 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
343 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
344 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
345 2-byte code at the end of the subject (it thought there wasn't enough data
346 left).
347
348 5. Comprehensive information about what went wrong is now returned by
349 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
350 as the output vector has at least 2 elements. The offset of the start of
351 the failing character and a reason code are placed in the vector.
352
353 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
354 now returned is for the first byte of the failing character, instead of the
355 last byte inspected. This is an incompatible change, but I hope it is small
356 enough not to be a problem. It makes the returned offset consistent with
357 pcre_exec() and pcre_dfa_exec().
358
359 7. pcretest now gives a text phrase as well as the error number when
360 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
361 failure, the offset and reason code are output.
362
363 8. When \R was used with a maximizing quantifier it failed to skip backwards
364 over a \r\n pair if the subsequent match failed. Instead, it just skipped
365 back over a single character (\n). This seems wrong (because it treated the
366 two characters as a single entity when going forwards), conflicts with the
367 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
368 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
369 has been changed.
370
371 9. Some internal refactoring has changed the processing so that the handling
372 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
373 time (the PCRE_DOTALL option was changed this way some time ago: version
374 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
375 which was always a bit of a fudge. It also means that there is one less
376 argument for the match() function, which reduces its stack requirements
377 slightly. This change also fixes an incompatibility with Perl: the pattern
378 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
379
380 10. More internal refactoring has drastically reduced the number of recursive
381 calls to match() for possessively repeated groups such as (abc)++ when
382 using pcre_exec().
383
384 11. While implementing 10, a number of bugs in the handling of groups were
385 discovered and fixed:
386
387 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
388 (a|)*(?1) gave a compile-time internal error.
389 ((a|)+)+ did not notice that the outer group could match an empty string.
390 (^a|^)+ was not marked as anchored.
391 (.*a|.*)+ was not marked as matching at start or after a newline.
392
393 12. Yet more internal refactoring has removed another argument from the match()
394 function. Special calls to this function are now indicated by setting a
395 value in a variable in the "match data" data block.
396
397 13. Be more explicit in pcre_study() instead of relying on "default" for
398 opcodes that mean there is no starting character; this means that when new
399 ones are added and accidentally left out of pcre_study(), testing should
400 pick them up.
401
402 14. The -s option of pcretest has been documented for ages as being an old
403 synonym of -m (show memory usage). I have changed it to mean "force study
404 for every regex", that is, assume /S for every regex. This is similar to -i
405 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
406 using it. It makes it easier to run collections of tests with and without
407 study enabled, and thereby test pcre_study() more easily. All the standard
408 tests are now run with and without -s (but some patterns can be marked as
409 "never study" - see 20 below).
410
411 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
412 restoration of the capturing data to the outer values was not happening
413 correctly.
414
415 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
416 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
417 pattern had matched an empty string, and so incorrectly returned a no
418 match.
419
420 17. There was optimizing code for the last branch of non-capturing parentheses,
421 and also for the obeyed branch of a conditional subexpression, which used
422 tail recursion to cut down on stack usage. Unfortunately, now that there is
423 the possibility of (*THEN) occurring in these branches, tail recursion is
424 no longer possible because the return has to be checked for (*THEN). These
425 two optimizations have therefore been removed. [But see 8.20/11 above.]
426
427 18. If a pattern containing \R was studied, it was assumed that \R always
428 matched two bytes, thus causing the minimum subject length to be
429 incorrectly computed because \R can also match just one byte.
430
431 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
432 was incorrectly computed.
433
434 20. If /S is present twice on a test pattern in pcretest input, it now
435 *disables* studying, thereby overriding the use of -s on the command line
436 (see 14 above). This is necessary for one or two tests to keep the output
437 identical in both cases.
438
439 21. When (*ACCEPT) was used in an assertion that matched an empty string and
440 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
441
442 22. When an atomic group that contained a capturing parenthesis was
443 successfully matched, but the branch in which it appeared failed, the
444 capturing was not being forgotten if a higher numbered group was later
445 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
446 group 1 to "a", when in fact it should be unset. This applied to multi-
447 branched capturing and non-capturing groups, repeated or not, and also to
448 positive assertions (capturing in negative assertions does not happen
449 in PCRE) and also to nested atomic groups.
450
451 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
452 subject after a captured substring, to make it easier to tell which of a
453 number of identical substrings has been captured.
454
455 24. The way atomic groups are processed by pcre_exec() has been changed so that
456 if they are repeated, backtracking one repetition now resets captured
457 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
458 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
459 "aaa". Previously, it would have been "a". As part of this code
460 refactoring, the way recursive calls are handled has also been changed.
461
462 25. If an assertion condition captured any substrings, they were not passed
463 back unless some other capturing happened later. For example, if
464 (?(?=(a))a) was matched against "a", no capturing was returned.
465
466 26. When studying a pattern that contained subroutine calls or assertions,
467 the code for finding the minimum length of a possible match was handling
468 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
469 group 1 called group 2 while simultaneously a separate group 2 called group
470 1). A stack overflow occurred in this case. I have fixed this by limiting
471 the recursion depth to 10.
472
473 27. Updated RunTest.bat in the distribution to the version supplied by Tom
474 Fortmann. This supports explicit test numbers on the command line, and has
475 argument validation and error reporting.
476
477 28. An instance of \X with an unlimited repeat could fail if at any point the
478 first character it looked at was a mark character.
479
480 29. Some minor code refactoring concerning Unicode properties and scripts
481 should reduce the stack requirement of match() slightly.
482
483 30. Added the '=' option to pcretest to check the setting of unused capturing
484 slots at the end of the pattern, which are documented as being -1, but are
485 not included in the return count.
486
487 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
488 compiled something random. Now it gives a compile-time error (as does
489 Perl).
490
491 32. A *MARK encountered during the processing of a positive assertion is now
492 recorded and passed back (compatible with Perl).
493
494 33. If --only-matching or --colour was set on a pcregrep call whose pattern
495 had alternative anchored branches, the search for a second match in a line
496 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
497 matched the line "0102" twice. The same bug affected patterns that started
498 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
499 twice.
500
501 34. Previously, PCRE did not allow quantification of assertions. However, Perl
502 does, and because of capturing effects, quantifying parenthesized
503 assertions may at times be useful. Quantifiers are now allowed for
504 parenthesized assertions.
505
506 35. A minor code tidy in pcre_compile() when checking options for \R usage.
507
508 36. \g was being checked for fancy things in a character class, when it should
509 just be a literal "g".
510
511 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
512 appearance of a nested POSIX class supersedes an apparent external class.
513 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
514 unescaped square brackets may also appear as part of class names. For
515 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
516 more like Perl. (But see 8.20/1 above.)
517
518 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
519 was because it thought it was \N{name}, which is not supported).
520
521 39. Add minix to OS list not supporting the -S option in pcretest.
522
523 40. PCRE tries to detect cases of infinite recursion at compile time, but it
524 cannot analyze patterns in sufficient detail to catch mutual recursions
525 such as ((?1))((?2)). There is now a runtime test that gives an error if a
526 subgroup is called recursively as a subpattern for a second time at the
527 same position in the subject string. In previous releases this might have
528 been caught by the recursion limit, or it might have run out of stack.
529
530 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
531 happen only once. PCRE was, however incorrectly giving a compile time error
532 "recursive call could loop indefinitely" because it cannot analyze the
533 pattern in sufficient detail. The compile time test no longer happens when
534 PCRE is compiling a conditional subpattern, but actual runaway loops are
535 now caught at runtime (see 40 above).
536
537 42. It seems that Perl allows any characters other than a closing parenthesis
538 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
539 has been changed to be the same.
540
541 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
542 as not to get warnings when autogen.sh is called. Also changed
543 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
544
545 44. To help people who use pcregrep to scan files containing exceedingly long
546 lines, the following changes have been made:
547
548 (a) The default value of the buffer size parameter has been increased from
549 8K to 20K. (The actual buffer used is three times this size.)
550
551 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
552 PCRE is built.
553
554 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
555 to be set at run time.
556
557 (d) Numerical values in pcregrep options can be followed by K or M, for
558 example --buffer-size=50K.
559
560 (e) If a line being scanned overflows pcregrep's buffer, an error is now
561 given and the return code is set to 2.
562
563 45. Add a pointer to the latest mark to the callout data block.
564
565 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
566 partial match of an empty string instead of no match. This was specific to
567 the use of ".".
568
569 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
570 complete match instead of a partial match. This bug was dependent on both
571 the PCRE_UTF8 and PCRE_DOTALL options being set.
572
573 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
574 starting byte set, because \b was not being ignored.
575
576
577 Version 8.12 15-Jan-2011
578 ------------------------
579
580 1. Fixed some typos in the markup of the man pages, and wrote a script that
581 checks for such things as part of the documentation building process.
582
583 2. On a big-endian 64-bit system, pcregrep did not correctly process the
584 --match-limit and --recursion-limit options (added for 8.11). In
585 particular, this made one of the standard tests fail. (The integer value
586 went into the wrong half of a long int.)
587
588 3. If the --colour option was given to pcregrep with -v (invert match), it
589 did strange things, either producing crazy output, or crashing. It should,
590 of course, ignore a request for colour when reporting lines that do not
591 match.
592
593 4. Another pcregrep bug caused similar problems if --colour was specified with
594 -M (multiline) and the pattern match finished with a line ending.
595
596 5. In pcregrep, when a pattern that ended with a literal newline sequence was
597 matched in multiline mode, the following line was shown as part of the
598 match. This seems wrong, so I have changed it.
599
600 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
601 the check for further matches in the same line (so they could be coloured)
602 to overrun the end of the current line. If another match was found, it was
603 incorrectly shown (and then shown again when found in the next line).
604
605 7. If pcregrep was compiled under Windows, there was a reference to the
606 function pcregrep_exit() before it was defined. I am assuming this was
607 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
608 reported by a user. I've moved the definition above the reference.
609
610
611 Version 8.11 10-Dec-2010
612 ------------------------
613
614 1. (*THEN) was not working properly if there were untried alternatives prior
615 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
616 backtracked to try for "b" instead of moving to the next alternative branch
617 at the same level (in this case, to look for "c"). The Perl documentation
618 is clear that when (*THEN) is backtracked onto, it goes to the "next
619 alternative in the innermost enclosing group".
620
621 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
622 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
623 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
624 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
625 (*THEN).
626
627 3. If \s appeared in a character class, it removed the VT character from
628 the class, even if it had been included by some previous item, for example
629 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
630 of \s, but is part of the POSIX "space" class.)
631
632 4. A partial match never returns an empty string (because you can always
633 match an empty string at the end of the subject); however the checking for
634 an empty string was starting at the "start of match" point. This has been
635 changed to the "earliest inspected character" point, because the returned
636 data for a partial match starts at this character. This means that, for
637 example, /(?<=abc)def/ gives a partial match for the subject "abc"
638 (previously it gave "no match").
639
640 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
641 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
642 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
643 has an implication that the given string is incomplete (because a partial
644 match is preferred over a full match). For this reason, these items now
645 give a partial match in this situation. [Aside: previously, the one case
646 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
647 match rather than a full match, which was wrong by the old rules, but is
648 now correct.]
649
650 6. There was a bug in the handling of #-introduced comments, recognized when
651 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
652 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
653 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
654 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
655 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
656 places in pcre_compile().
657
658 7. Related to (6) above, when pcre_compile() was skipping #-introduced
659 comments when looking ahead for named forward references to subpatterns,
660 the only newline sequence it recognized was NL. It now handles newlines
661 according to the set newline convention.
662
663 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
664 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
665 cater for a lack of strerror(). These oversights have been fixed.
666
667 9. Added --match-limit and --recursion-limit to pcregrep.
668
669 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
670
671 11. When the -o option was used, pcregrep was setting a return code of 1, even
672 when matches were found, and --line-buffered was not being honoured.
673
674 12. Added an optional parentheses number to the -o and --only-matching options
675 of pcregrep.
676
677 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
678 can match an empty string. The code to do it in pcretest and pcredemo
679 needed fixing:
680
681 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
682 only one byte after an empty string match just before CRLF (this case
683 just got forgotten; "any" and "anycrlf" were OK).
684
685 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
686 mode when an empty string match preceded an ASCII character followed by
687 a non-ASCII character. (The code for advancing by one character rather
688 than one byte was nonsense.)
689
690 (c) The pcredemo.c sample program did not have any code at all to handle
691 the cases when CRLF is a valid newline sequence.
692
693 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
694 as a starting offset was within the subject string. There is now a new
695 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
696 negative or greater than the length of the string. In order to test this,
697 pcretest is extended to allow the setting of negative starting offsets.
698
699 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
700 starting offset points to the beginning of a UTF-8 character was
701 unnecessarily clumsy. I tidied it up.
702
703 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
704 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
705
706 17. Nobody had reported that the --include_dir option, which was added in
707 release 7.7 should have been called --include-dir (hyphen, not underscore)
708 for compatibility with GNU grep. I have changed it to --include-dir, but
709 left --include_dir as an undocumented synonym, and the same for
710 --exclude-dir, though that is not available in GNU grep, at least as of
711 release 2.5.4.
712
713 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
714 characters from a string of bytes) have been redefined so as not to use
715 loops, in order to improve performance in some environments. At the same
716 time, I abstracted some of the common code into auxiliary macros to save
717 repetition (this should not affect the compiled code).
718
719 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
720 compile-time error is now given if \c is not followed by an ASCII
721 character, that is, a byte less than 128. (In EBCDIC mode, the code is
722 different, and any byte value is allowed.)
723
724 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
725 START_OPTIMIZE option, which is now allowed at compile time - but just
726 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
727 to pcregrep and other applications that have no direct access to PCRE
728 options. The new /Y option in pcretest sets this option when calling
729 pcre_compile().
730
731 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
732 back references. Groups containing recursive back references were forced to
733 be atomic by that change, but in the case of named groups, the amount of
734 memory required was incorrectly computed, leading to "Failed: internal
735 error: code overflow". This has been fixed.
736
737 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
738 pcretest.c, to avoid build problems in some Borland environments.
739
740
741 Version 8.10 25-Jun-2010
742 ------------------------
743
744 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
745 THEN.
746
747 2. (*ACCEPT) was not working when inside an atomic group.
748
749 3. Inside a character class, \B is treated as a literal by default, but
750 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
751 causes the error). The code is unchanged, but I tidied the documentation.
752
753 4. Inside a character class, PCRE always treated \R and \X as literals,
754 whereas Perl faults them if its -w option is set. I have changed PCRE so
755 that it faults them when PCRE_EXTRA is set.
756
757 5. Added support for \N, which always matches any character other than
758 newline. (It is the same as "." when PCRE_DOTALL is not set.)
759
760 6. When compiling pcregrep with newer versions of gcc which may have
761 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
762 declared with attribute warn_unused_result" were given. Just casting the
763 result to (void) does not stop the warnings; a more elaborate fudge is
764 needed. I've used a macro to implement this.
765
766 7. Minor change to pcretest.c to avoid a compiler warning.
767
768 8. Added four artifical Unicode properties to help with an option to make
769 \s etc use properties (see next item). The new properties are: Xan
770 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
771
772 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
773 use Unicode properties. (*UCP) at the start of a pattern can be used to set
774 this option. Modified pcretest to add /W to test this facility. Added
775 REG_UCP to make it available via the POSIX interface.
776
777 10. Added --line-buffered to pcregrep.
778
779 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
780 studied, and the match started with a letter with a code point greater than
781 127 whose first byte was different to the first byte of the other case of
782 the letter, the other case of this starting letter was not recognized
783 (#976).
784
785 12. If a pattern that was studied started with a repeated Unicode property
786 test, for example, \p{Nd}+, there was the theoretical possibility of
787 setting up an incorrect bitmap of starting bytes, but fortunately it could
788 not have actually happened in practice until change 8 above was made (it
789 added property types that matched character-matching opcodes).
790
791 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
792 possible starting bytes for non-anchored patterns.
793
794 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
795 \R, and also a number of cases that involve Unicode properties, both
796 explicit and implicit when PCRE_UCP is set.
797
798 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
799 input, it could crash or give wrong results if characters with values
800 greater than 0xc0 were present in the subject string. (Detail: it assumed
801 UTF-8 input when processing these items.)
802
803 16. Added a lot of (int) casts to avoid compiler warnings in systems where
804 size_t is 64-bit (#991).
805
806 17. Added a check for running out of memory when PCRE is compiled with
807 --disable-stack-for-recursion (#990).
808
809 18. If the last data line in a file for pcretest does not have a newline on
810 the end, a newline was missing in the output.
811
812 19. The default pcre_chartables.c file recognizes only ASCII characters (values
813 less than 128) in its various bitmaps. However, there is a facility for
814 generating tables according to the current locale when PCRE is compiled. It
815 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
816 characters, are recognized by isspace() and therefore were getting set in
817 these tables, and indeed these tables seem to approximate to ISO 8859. This
818 caused a problem in UTF-8 mode when pcre_study() was used to create a list
819 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
820 which of course cannot start UTF-8 characters. I have changed the code so
821 that only real ASCII characters (less than 128) and the correct starting
822 bytes for UTF-8 encodings are set for characters greater than 127 when in
823 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
824 altogether.)
825
826 20. Added the /T option to pcretest so as to be able to run tests with non-
827 standard character tables, thus making it possible to include the tests
828 used for 19 above in the standard set of tests.
829
830 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
831 reference to a subpattern the other side of a comment that contains an
832 opening parenthesis caused either an internal compiling error, or a
833 reference to the wrong subpattern.
834
835
836 Version 8.02 19-Mar-2010
837 ------------------------
838
839 1. The Unicode data tables have been updated to Unicode 5.2.0.
840
841 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
842 configured.
843
844 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
845 original author of that file, following a query about its status.
846
847 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
848 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
849
850 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
851 quantifier applied to a forward-referencing subroutine call, could compile
852 incorrect code or give the error "internal error: previously-checked
853 referenced subpattern not found".
854
855 6. Both MS Visual Studio and Symbian OS have problems with initializing
856 variables to point to external functions. For these systems, therefore,
857 pcre_malloc etc. are now initialized to local functions that call the
858 relevant global functions.
859
860 7. There were two entries missing in the vectors called coptable and poptable
861 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
862 I've fixed the data, and added a kludgy way of testing at compile time that
863 the lengths are correct (equal to the number of opcodes).
864
865 8. Following on from 7, I added a similar kludge to check the length of the
866 eint vector in pcreposix.c.
867
868 9. Error texts for pcre_compile() are held as one long string to avoid too
869 much relocation at load time. To find a text, the string is searched,
870 counting zeros. There was no check for running off the end of the string,
871 which could happen if a new error number was added without updating the
872 string.
873
874 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
875
876 11. \K was not working if it appeared in an atomic group or in a group that
877 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
878 \K is "not well defined" if used in an assertion. PCRE now accepts it if
879 the assertion is positive, but not if it is negative.
880
881 12. Change 11 fortuitously reduced the size of the stack frame used in the
882 "match()" function of pcre_exec.c by one pointer. Forthcoming
883 implementation of support for (*MARK) will need an extra pointer on the
884 stack; I have reserved it now, so that the stack frame size does not
885 decrease.
886
887 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
888 item in branch that calls a recursion is a subroutine call - as in the
889 second branch in the above example - was incorrectly given the compile-
890 time error "recursive call could loop indefinitely" because pcre_compile()
891 was not correctly checking the subroutine for matching a non-empty string.
892
893 14. The checks for overrunning compiling workspace could trigger after an
894 overrun had occurred. This is a "should never occur" error, but it can be
895 triggered by pathological patterns such as hundreds of nested parentheses.
896 The checks now trigger 100 bytes before the end of the workspace.
897
898 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
899
900
901 Version 8.01 19-Jan-2010
902 ------------------------
903
904 1. If a pattern contained a conditional subpattern with only one branch (in
905 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
906 computed the wrong minimum data length (which is of course zero for such
907 subpatterns). This could cause incorrect "no match" results.
908
909 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
910 the pattern is reset in the first branch, pcre_compile() failed with
911 "internal error: code overflow at offset...". This happened only when
912 the reset was to the original external option setting. (An optimization
913 abstracts leading options settings into an external setting, which was the
914 cause of this.)
915
916 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
917 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
918 assertion pattern did not match (meaning that the assertion was true), it
919 was incorrectly treated as false if the SKIP had been reached during the
920 matching. This also applied to assertions used as conditions.
921
922 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
923 assertion subpattern, including such a pattern used as a condition,
924 unpredictable results occurred, instead of the error return
925 PCRE_ERROR_DFA_UITEM.
926
927 5. The C++ GlobalReplace function was not working like Perl for the special
928 situation when an empty string is matched. It now does the fancy magic
929 stuff that is necessary.
930
931 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
932 removed. (These were left over from very, very early versions of PCRE.)
933
934 7. Some cosmetic changes to the code to make life easier when compiling it
935 as part of something else:
936
937 (a) Change DEBUG to PCRE_DEBUG.
938
939 (b) In pcre_compile(), rename the member of the "branch_chain" structure
940 called "current" as "current_branch", to prevent a collision with the
941 Linux macro when compiled as a kernel module.
942
943 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
944 prevent a collision with the Linux macro when compiled as a kernel
945 module.
946
947 8. In pcre_compile() there are some checks for integer overflows that used to
948 cast potentially large values to (double). This has been changed to that
949 when building, a check for int64_t is made, and if it is found, it is used
950 instead, thus avoiding the use of floating point arithmetic. (There is no
951 other use of FP in PCRE.) If int64_t is not found, the fallback is to
952 double.
953
954 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
955 2005 (difference between two addresses compared to an unsigned value).
956
957 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
958 custom one, because of the following reported problem in Windows:
959
960 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
961 under Win32.
962 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
963 therefore missing the function definition.
964 - The compiler thus generates a "C" signature for the test function.
965 - The linker fails to find the "C" function.
966 - PCRE fails to configure if asked to do so against libbz2.
967
968 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
969 messages were output:
970
971 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
972 rerunning libtoolize, to keep the correct libtool macros in-tree.
973 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
974
975 I have done both of these things.
976
977 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
978 most of the time, it *can* run out if it is given a pattern that contains a
979 runaway infinite recursion. I updated the discussion in the pcrestack man
980 page.
981
982 13. Now that we have gone to the x.xx style of version numbers, the minor
983 version may start with zero. Using 08 or 09 is a bad idea because users
984 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
985 interpreted as invalid octal numbers. I've updated the previous comment in
986 configure.ac, and also added a check that gives an error if 08 or 09 are
987 used.
988
989 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
990 causing partial matching to fail when the end of the subject matched \W
991 in a UTF-8 pattern where \W was quantified with a minimum of 3.
992
993 15. There were some discrepancies between the declarations in pcre_internal.h
994 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
995 their definitions. The declarations used "const uschar *" and the
996 definitions used USPTR. Even though USPTR is normally defined as "const
997 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
998 reported that: "This difference in casting confuses some C++ compilers, for
999 example, SunCC recognizes above declarations as different functions and
1000 generates broken code for hbpcre." I have changed the declarations to use
1001 USPTR.
1002
1003 16. GNU libtool is named differently on some systems. The autogen.sh script now
1004 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1005 (FreeBSD).
1006
1007 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1008 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1009 comment: "Figure out how to create a longlong from a string: strtoll and
1010 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1011 instance, but it only takes 2 args instead of 3!"
1012
1013 18. A subtle bug concerned with back references has been fixed by a change of
1014 specification, with a corresponding code fix. A pattern such as
1015 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1016 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1017 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1018 same bug. Such groups have to be quantified to be useful, or contained
1019 inside another quantified group. (If there's no repetition, the reference
1020 can never match.) The problem arises because, having left the group and
1021 moved on to the rest of the pattern, a later failure that backtracks into
1022 the group uses the captured value from the final iteration of the group
1023 rather than the correct earlier one. I have fixed this in PCRE by forcing
1024 any group that contains a reference to itself to be an atomic group; that
1025 is, there cannot be any backtracking into it once it has completed. This is
1026 similar to recursive and subroutine calls.
1027
1028
1029 Version 8.00 19-Oct-09
1030 ----------------------
1031
1032 1. The table for translating pcre_compile() error codes into POSIX error codes
1033 was out-of-date, and there was no check on the pcre_compile() error code
1034 being within the table. This could lead to an OK return being given in
1035 error.
1036
1037 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1038 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1039 in a Windows environment.
1040
1041 3. The pcregrep --count option prints the count for each file even when it is
1042 zero, as does GNU grep. However, pcregrep was also printing all files when
1043 --files-with-matches was added. Now, when both options are given, it prints
1044 counts only for those files that have at least one match. (GNU grep just
1045 prints the file name in this circumstance, but including the count seems
1046 more useful - otherwise, why use --count?) Also ensured that the
1047 combination -clh just lists non-zero counts, with no names.
1048
1049 4. The long form of the pcregrep -F option was incorrectly implemented as
1050 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1051 but it seems right to fix it, and I didn't think it was worth preserving
1052 the old behaviour.
1053
1054 5. The command line items --regex=pattern and --regexp=pattern were not
1055 recognized by pcregrep, which required --regex pattern or --regexp pattern
1056 (with a space rather than an '='). The man page documented the '=' forms,
1057 which are compatible with GNU grep; these now work.
1058
1059 6. No libpcreposix.pc file was created for pkg-config; there was just
1060 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1061
1062 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1063 when UCP support is not needed, by modifying the Python script that
1064 generates it from Unicode data files. This should not matter if the module
1065 is correctly used as a library, but I received one complaint about 50K of
1066 unwanted data. My guess is that the person linked everything into his
1067 program rather than using a library. Anyway, it does no harm.
1068
1069 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1070 was a minimum greater than 1 for a wide character in a possessive
1071 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1072 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1073 character. Chaos in the form of incorrect output or a compiling loop could
1074 result.
1075
1076 9. The restrictions on what a pattern can contain when partial matching is
1077 requested for pcre_exec() have been removed. All patterns can now be
1078 partially matched by this function. In addition, if there are at least two
1079 slots in the offset vector, the offset of the earliest inspected character
1080 for the match and the offset of the end of the subject are set in them when
1081 PCRE_ERROR_PARTIAL is returned.
1082
1083 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1084 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1085 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1086 and may be more useful for multi-segment matching.
1087
1088 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1089 used to be given if ever the end of the subject was reached; now it is
1090 given only if matching could not proceed because another character was
1091 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1092 string "Z", which now yields "no match" instead of "partial match". In the
1093 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1094 final character ended with (*FAIL).
1095
1096 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1097 if the pattern had a "must contain" character that was already found in the
1098 earlier partial match, unless partial matching was again requested. For
1099 example, with the pattern /dog.(body)?/, the "must contain" character is
1100 "g". If the first part-match was for the string "dog", restarting with
1101 "sbody" failed. This bug has been fixed.
1102
1103 13. The string returned by pcre_dfa_exec() after a partial match has been
1104 changed so that it starts at the first inspected character rather than the
1105 first character of the match. This makes a difference only if the pattern
1106 starts with a lookbehind assertion or \b or \B (\K is not supported by
1107 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1108 matching functions compatible, and I think it's the right thing to do.
1109
1110 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1111 so that the demonstration program is easily available in environments where
1112 PCRE has not been installed from source.
1113
1114 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1115 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1116 library.
1117
1118 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1119 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1120 is not the first non-POSIX option to be added. Clearly some people find
1121 these options useful.
1122
1123 17. If a caller to the POSIX matching function regexec() passes a non-zero
1124 value for nmatch with a NULL value for pmatch, the value of
1125 nmatch is forced to zero.
1126
1127 18. RunGrepTest did not have a test for the availability of the -u option of
1128 the diff command, as RunTest does. It now checks in the same way as
1129 RunTest, and also checks for the -b option.
1130
1131 19. If an odd number of negated classes containing just a single character
1132 interposed, within parentheses, between a forward reference to a named
1133 subpattern and the definition of the subpattern, compilation crashed with
1134 an internal error, complaining that it could not find the referenced
1135 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1136 [The bug was that it was starting one character too far in when skipping
1137 over the character class, thus treating the ] as data rather than
1138 terminating the class. This meant it could skip too much.]
1139
1140 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1141 /g option in pcretest when the pattern contains \K, which makes it possible
1142 to have an empty string match not at the start, even when the pattern is
1143 anchored. Updated pcretest and pcredemo to use this option.
1144
1145 21. If the maximum number of capturing subpatterns in a recursion was greater
1146 than the maximum at the outer level, the higher number was returned, but
1147 with unset values at the outer level. The correct (outer level) value is
1148 now given.
1149
1150 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1151 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1152 make it do so. The string so far is captured, making this feature
1153 compatible with Perl.
1154
1155 23. The tests have been re-organized, adding tests 11 and 12, to make it
1156 possible to check the Perl 5.10 features against Perl 5.10.
1157
1158 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1159 pattern matches a fixed length string. PCRE did not allow this; now it
1160 does. Neither allows recursion.
1161
1162 25. I finally figured out how to implement a request to provide the minimum
1163 length of subject string that was needed in order to match a given pattern.
1164 (It was back references and recursion that I had previously got hung up
1165 on.) This code has now been added to pcre_study(); it finds a lower bound
1166 to the length of subject needed. It is not necessarily the greatest lower
1167 bound, but using it to avoid searching strings that are too short does give
1168 some useful speed-ups. The value is available to calling programs via
1169 pcre_fullinfo().
1170
1171 26. While implementing 25, I discovered to my embarrassment that pcretest had
1172 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1173 study optimizations had never been tested with that matching function.
1174 Oops. What is worse, even when it was passed study data, there was a bug in
1175 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1176 were also very few tests of studied patterns with pcre_dfa_exec().
1177
1178 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1179 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1180 on the other side of the coin, they are no longer allowed to have different
1181 names, because these cannot be distinguished in PCRE, and this has caused
1182 confusion. (This is a difference from Perl.)
1183
1184 28. When duplicate subpattern names are present (necessarily with different
1185 numbers, as required by 27 above), and a test is made by name in a
1186 conditional pattern, either for a subpattern having been matched, or for
1187 recursion in such a pattern, all the associated numbered subpatterns are
1188 tested, and the overall condition is true if the condition is true for any
1189 one of them. This is the way Perl works, and is also more like the way
1190 testing by number works.
1191
1192
1193 Version 7.9 11-Apr-09
1194 ---------------------
1195
1196 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1197 (pcretest), all targets were linked against these libraries. This included
1198 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1199 libraries. This caused unwanted dependencies to be created. This problem
1200 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1201 pcretest is linked with readline.
1202
1203 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1204 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1205 moved outside it again, because FALSE and TRUE are already defined in AIX,
1206 but BOOL is not.
1207
1208 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1209 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1210
1211 4. The pcregrep documentation said spaces were inserted as well as colons (or
1212 hyphens) following file names and line numbers when outputting matching
1213 lines. This is not true; no spaces are inserted. I have also clarified the
1214 wording for the --colour (or --color) option.
1215
1216 5. In pcregrep, when --colour was used with -o, the list of matching strings
1217 was not coloured; this is different to GNU grep, so I have changed it to be
1218 the same.
1219
1220 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1221 each matching line was coloured. Now it goes on to look for further matches
1222 of any of the test patterns, which is the same behaviour as GNU grep.
1223
1224 7. A pattern that could match an empty string could cause pcregrep to loop; it
1225 doesn't make sense to accept an empty string match in pcregrep, so I have
1226 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1227 seems to be how GNU grep behaves.
1228
1229 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1230 start or after a newline", because the conditional assertion was not being
1231 correctly handled. The rule now is that both the assertion and what follows
1232 in the first alternative must satisfy the test.
1233
1234 9. If auto-callout was enabled in a pattern with a conditional group whose
1235 condition was an assertion, PCRE could crash during matching, both with
1236 pcre_exec() and pcre_dfa_exec().
1237
1238 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1239 used for matching.
1240
1241 11. Unicode property support in character classes was not working for
1242 characters (bytes) greater than 127 when not in UTF-8 mode.
1243
1244 12. Added the -M command line option to pcretest.
1245
1246 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1247
1248 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1249
1250 16. Added comments and documentation about mis-use of no_arg in the C++
1251 wrapper.
1252
1253 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1254 from Martin Jerabek that uses macro names for all relevant character and
1255 string constants.
1256
1257 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1258 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1259 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1260 these, but not everybody uses configure.
1261
1262 19. A conditional group that had only one branch was not being correctly
1263 recognized as an item that could match an empty string. This meant that an
1264 enclosing group might also not be so recognized, causing infinite looping
1265 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1266 with the subject "ab", where knowledge that the repeated group can match
1267 nothing is needed in order to break the loop.
1268
1269 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1270 exec(), but without supplying a callout function, matching went wrong.
1271
1272 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1273 leak if the size of the offset vector was greater than 30. When the vector
1274 is smaller, the saved offsets during recursion go onto a local stack
1275 vector, but for larger vectors malloc() is used. It was failing to free
1276 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1277 error, in fact).
1278
1279 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1280 heapframe that is used only when UTF-8 support is enabled. This caused no
1281 problem, but was untidy.
1282
1283 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1284 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1285 included within another project.
1286
1287 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1288 slightly modified by me:
1289
1290 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1291 not building pcregrep.
1292
1293 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1294 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1295
1296 25. Forward references, both numeric and by name, in patterns that made use of
1297 duplicate group numbers, could behave incorrectly or give incorrect errors,
1298 because when scanning forward to find the reference group, PCRE was not
1299 taking into account the duplicate group numbers. A pattern such as
1300 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1301
1302 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1303 the feature of a custom pointer more persuasive (as requested by a user).
1304
1305 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1306 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1307 pre-defined.
1308
1309 28. Added support for (*UTF8) at the start of a pattern.
1310
1311 29. Arrange for flags added by the "release type" setting in CMake to be shown
1312 in the configuration summary.
1313
1314
1315 Version 7.8 05-Sep-08
1316 ---------------------
1317
1318 1. Replaced UCP searching code with optimized version as implemented for Ad
1319 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1320 stage table and inline lookup instead of a function, giving speed ups of 2
1321 to 5 times on some simple patterns that I tested. Permission was given to
1322 distribute the MultiStage2.py script that generates the tables (it's not in
1323 the tarball, but is in the Subversion repository).
1324
1325 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1326 scripts.
1327
1328 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1329 a group with a zero qualifier. The result of the study could be incorrect,
1330 or the function might crash, depending on the pattern.
1331
1332 4. Caseless matching was not working for non-ASCII characters in back
1333 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1334 It now works when Unicode Property Support is available.
1335
1336 5. In pcretest, an escape such as \x{de} in the data was always generating
1337 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1338 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1339 truncation.
1340
1341 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1342
1343 7. Added two (int) casts to pcregrep when printing the difference of two
1344 pointers, in case they are 64-bit values.
1345
1346 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1347 test 2 if it fails.
1348
1349 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1350 and a #define of that name to empty if it is not externally set. This is to
1351 allow users of MSVC to set it if necessary.
1352
1353 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1354 the convenience functions in the pcre_get.c source file.
1355
1356 11. An option change at the start of a pattern that had top-level alternatives
1357 could cause overwriting and/or a crash. This command provoked a crash in
1358 some environments:
1359
1360 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1361
1362 This potential security problem was recorded as CVE-2008-2371.
1363
1364 12. For a pattern where the match had to start at the beginning or immediately
1365 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1366 pcre_dfa_exec() could read past the end of the passed subject if there was
1367 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1368 pcretest so that it places the subject at the end of its malloc-ed buffer.
1369
1370 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1371 exec() might read past the end of the data buffer in UTF-8 mode.
1372
1373 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1374 the data contained the byte 0x85 as part of a UTF-8 character within its
1375 first line. This applied both to normal and DFA matching.
1376
1377 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1378 /^[^d]*?$/8 failed to match "abc".
1379
1380 16. Added a missing copyright notice to pcrecpp_internal.h.
1381
1382 17. Make it more clear in the documentation that values returned from
1383 pcre_exec() in ovector are byte offsets, not character counts.
1384
1385 18. Tidied a few places to stop certain compilers from issuing warnings.
1386
1387 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1388 supplied by Stefan Weber. I made a further small update for 7.8 because
1389 there is a change of source arrangements: the pcre_searchfuncs.c module is
1390 replaced by pcre_ucd.c.
1391
1392
1393 Version 7.7 07-May-08
1394 ---------------------
1395
1396 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1397 a string to a long long, pretend we don't even have a long long." This is
1398 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1399
1400 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1401 pre-7.6 versions, which defined a global no_arg variable instead of putting
1402 it in the RE class. (See also #8 below.)
1403
1404 3. Remove a line of dead code, identified by coverity and reported by Nuno
1405 Lopes.
1406
1407 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1408
1409 (1) The include/exclude patterns were being applied to the whole pathnames
1410 of files, instead of just to the final components.
1411
1412 (2) If there was more than one level of directory, the subdirectories were
1413 skipped unless they satisfied the include/exclude conditions. This is
1414 inconsistent with GNU grep (and could even be seen as contrary to the
1415 pcregrep specification - which I improved to make it absolutely clear).
1416 The action now is always to scan all levels of directory, and just
1417 apply the include/exclude patterns to regular files.
1418
1419 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1420 --exclude_dir in the tests to avoid scanning .svn directories.
1421
1422 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1423 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1424 doesn't support NULs in patterns.
1425
1426 7. Added some missing "const"s to declarations of static tables in
1427 pcre_compile.c and pcre_dfa_exec.c.
1428
1429 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1430 caused by fix #2 above. (Subsequently also a second patch to fix the
1431 first patch. And a third patch - this was a messy problem.)
1432
1433 9. Applied Craig's patch to remove the use of push_back().
1434
1435 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1436 matching function regexec().
1437
1438 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1439 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1440 references. PCRE supports relative numbers with this syntax (I don't think
1441 Oniguruma does).
1442
1443 12. Previously, a group with a zero repeat such as (...){0} was completely
1444 omitted from the compiled regex. However, this means that if the group
1445 was called as a subroutine from elsewhere in the pattern, things went wrong
1446 (an internal error was given). Such groups are now left in the compiled
1447 pattern, with a new opcode that causes them to be skipped at execution
1448 time.
1449
1450 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1451 to the way PCRE behaves:
1452
1453 (a) A lone ] character is dis-allowed (Perl treats it as data).
1454
1455 (b) A back reference to an unmatched subpattern matches an empty string
1456 (Perl fails the current match path).
1457
1458 (c) A data ] in a character class must be notated as \] because if the
1459 first data character in a class is ], it defines an empty class. (In
1460 Perl it is not possible to have an empty class.) The empty class []
1461 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1462 The negative empty class [^] matches any one character, independently
1463 of the DOTALL setting.
1464
1465 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1466 non-existent subpattern following a character class starting with ']' and
1467 containing () gave an internal compiling error instead of "reference to
1468 non-existent subpattern". Fortunately, when the pattern did exist, the
1469 compiled code was correct. (When scanning forwards to check for the
1470 existencd of the subpattern, it was treating the data ']' as terminating
1471 the class, so got the count wrong. When actually compiling, the reference
1472 was subsequently set up correctly.)
1473
1474 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1475 it was being rejected as not supported by pcre_dfa_exec(), even though
1476 other assertions are supported. I have made pcre_dfa_exec() support
1477 (*FAIL).
1478
1479 16. The implementation of 13c above involved the invention of a new opcode,
1480 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1481 cannot be changed at match time, I realized I could make a small
1482 improvement to matching performance by compiling OP_ALLANY instead of
1483 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1484 on the OP_ANY path.
1485
1486 17. Compiling pcretest on Windows with readline support failed without the
1487 following two fixes: (1) Make the unistd.h include conditional on
1488 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1489
1490 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1491 ncurses library to be included for pcretest when ReadLine support is
1492 requested, but also to allow for it to be overridden. This patch came from
1493 Daniel Bergström.
1494
1495 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1496 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1497 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1498 spotting this.
1499
1500
1501 Version 7.6 28-Jan-08
1502 ---------------------
1503
1504 1. A character class containing a very large number of characters with
1505 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1506 overflow.
1507
1508 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1509 HAVE_LONG_LONG is not defined.
1510
1511 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1512 bring them up to date and include new features. This patch includes:
1513
1514 - Fixed PH's badly added libz and libbz2 support.
1515 - Fixed a problem with static linking.
1516 - Added pcredemo. [But later removed - see 7 below.]
1517 - Fixed dftables problem and added an option.
1518 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1519 HAVE_LONG_LONG.
1520 - Added readline support for pcretest.
1521 - Added an listing of the option settings after cmake has run.
1522
1523 4. A user submitted a patch to Makefile that makes it easy to create
1524 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1525 Makefile.am that cause it to include this special target, without
1526 affecting anything else. Note that the same mingw target plus all
1527 the other distribution libraries and programs are now supported
1528 when configuring with CMake (see 6 below) instead of with
1529 Configure/Make.
1530
1531 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1532 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1533 exported in the Windows port". It has not yet been confirmed that the patch
1534 solves the problem, but it does no harm.
1535
1536 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1537 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1538 with CMake, and also correct the comment about stack recursion.
1539
1540 7. Remove the automatic building of pcredemo from the ./configure system and
1541 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1542 of a program that users should build themselves after PCRE is installed, so
1543 building it automatically is not really right. What is more, it gave
1544 trouble in some build environments.
1545
1546 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1547
1548
1549 Version 7.5 10-Jan-08
1550 ---------------------
1551
1552 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1553 values in parens when parsing an RE using the C++ wrapper."
1554
1555 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1556 Characters greater than 255 were excluded from the class instead of being
1557 included.
1558
1559 3. The same bug as (2) above applied to negated POSIX classes such as
1560 [:^space:].
1561
1562 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1563 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1564 I have changed it.
1565
1566 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1567 first named subpattern) and a construct such as (?&a) would reference the
1568 first named subpattern whose name started with "a" (in other words, the
1569 length check was missing). Both these problems are fixed. "Subpattern name
1570 expected" is now given for (?&) (a zero-length name), and this patch also
1571 makes it give the same error for \k'' (previously it complained that that
1572 was a reference to a non-existent subpattern).
1573
1574 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1575 this is right because (?- can be followed by option settings as well as by
1576 digits. I have, however, made the messages clearer.
1577
1578 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1579 than the number used in the conditional) now cause a compile-time error.
1580 This is actually not compatible with Perl, which accepts such patterns, but
1581 treats the conditional as always being FALSE (as PCRE used to), but it
1582 seems to me that giving a diagnostic is better.
1583
1584 8. Change "alphameric" to the more common word "alphanumeric" in comments
1585 and messages.
1586
1587 9. Fix two occurrences of "backslash" in comments that should have been
1588 "backspace".
1589
1590 10. Remove two redundant lines of code that can never be obeyed (their function
1591 was moved elsewhere).
1592
1593 11. The program that makes PCRE's Unicode character property table had a bug
1594 which caused it to generate incorrect table entries for sequences of
1595 characters that have the same character type, but are in different scripts.
1596 It amalgamated them into a single range, with the script of the first of
1597 them. In other words, some characters were in the wrong script. There were
1598 thirteen such cases, affecting characters in the following ranges:
1599
1600 U+002b0 - U+002c1
1601 U+0060c - U+0060d
1602 U+0061e - U+00612
1603 U+0064b - U+0065e
1604 U+0074d - U+0076d
1605 U+01800 - U+01805
1606 U+01d00 - U+01d77
1607 U+01d9b - U+01dbf
1608 U+0200b - U+0200f
1609 U+030fc - U+030fe
1610 U+03260 - U+0327f
1611 U+0fb46 - U+0fbb1
1612 U+10450 - U+1049d
1613
1614 12. The -o option (show only the matching part of a line) for pcregrep was not
1615 compatible with GNU grep in that, if there was more than one match in a
1616 line, it showed only the first of them. It now behaves in the same way as
1617 GNU grep.
1618
1619 13. If the -o and -v options were combined for pcregrep, it printed a blank
1620 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1621 does the same. The return code can be used to tell if there were any
1622 non-matching lines.
1623
1624 14. Added --file-offsets and --line-offsets to pcregrep.
1625
1626 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1627 infinitely looping recursion. The bug was that positive lookaheads were not
1628 being skipped when checking for a possible empty match (negative lookaheads
1629 and both kinds of lookbehind were skipped).
1630
1631 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1632 inclusion of <windows.h> to before rather than after the definition of
1633 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1634
1635 17. Specifying a possessive quantifier with a specific limit for a Unicode
1636 character property caused pcre_compile() to compile bad code, which led at
1637 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1638 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1639 caused the error; without that there was no problem.
1640
1641 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1642
1643 19. Added --enable-pcretest-libreadline.
1644
1645 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1646 RE::GlobalReplace(). As a result, the number of replacements returned was
1647 double what it should be. I removed one of the increments, but Craig sent a
1648 later patch that removed the other one (the right fix) and added unit tests
1649 that check the return values (which was not done before).
1650
1651 21. Several CMake things:
1652
1653 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1654 the names libpcre and libpcreposix, not just pcre and pcreposix.
1655
1656 (2) The above change means that pcretest and pcregrep are now correctly
1657 linked with the newly-built libraries, not previously installed ones.
1658
1659 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1660
1661 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1662 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1663 UTF-8 newline character). The key issue is that the pattern starts .*;
1664 this means that the match must be either at the beginning, or after a
1665 newline. The bug was in the code for advancing after a failed match and
1666 checking that the new position followed a newline. It was not taking
1667 account of UTF-8 characters correctly.
1668
1669 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1670 character classes. PCRE was not treating the sequence [:...:] as a
1671 character class unless the ... were all letters. Perl, however, seems to
1672 allow any characters between [: and :], though of course it rejects as
1673 unknown any "names" that contain non-letters, because all the known class
1674 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1675 for example, whereas PCRE did not - it did not recognize a POSIX character
1676 class. This seemed a bit dangerous, so the code has been changed to be
1677 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1678 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1679 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1680 Perl does, and where it didn't before.
1681
1682 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1683 Windows environments %n is disabled by default.
1684
1685
1686 Version 7.4 21-Sep-07
1687 ---------------------
1688
1689 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1690 means that a class such as [\s] counted as "explicit reference to CR or
1691 LF". That isn't really right - the whole point of the change was to try to
1692 help when there was an actual mention of one of the two characters. So now
1693 the change happens only if \r or \n (or a literal CR or LF) character is
1694 encountered.
1695
1696 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1697 of both had grown to the point where there were only 3 bits left.
1698 Fortunately, there was spare space in the data structure, and so I have
1699 moved the internal flags into a new 16-bit field to free up more option
1700 bits.
1701
1702 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1703 but did not set the internal JCHANGED flag - either of these is enough to
1704 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1705 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1706 start sets both bits.
1707
1708 4. Added options (at build time, compile time, exec time) to change \R from
1709 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1710
1711 5. doc/pcresyntax.html was missing from the distribution.
1712
1713 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1714 compatibility, even though it is no longer used.
1715
1716 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1717 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1718 windows.h file is present (where different names are used). [This was
1719 reversed later after testing - see 16 below.]
1720
1721 8. Changed all #include <config.h> to #include "config.h". There were also
1722 some further <pcre.h> cases that I changed to "pcre.h".
1723
1724 9. When pcregrep was used with the --colour option, it missed the line ending
1725 sequence off the lines that it output.
1726
1727 10. It was pointed out to me that arrays of string pointers cause lots of
1728 relocations when a shared library is dynamically loaded. A technique of
1729 using a single long string with a table of offsets can drastically reduce
1730 these. I have refactored PCRE in four places to do this. The result is
1731 dramatic:
1732
1733 Originally: 290
1734 After changing UCP table: 187
1735 After changing error message table: 43
1736 After changing table of "verbs" 36
1737 After changing table of Posix names 22
1738
1739 Thanks to the folks working on Gregex for glib for this insight.
1740
1741 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1742 unicode-properties was also set.
1743
1744 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1745
1746 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1747 checked only for CRLF.
1748
1749 14. Added casts to pcretest.c to avoid compiler warnings.
1750
1751 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1752
1753 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1754 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1755 entirely. This removes changes made in 7 above.
1756
1757 17. The CMake files have been updated, and there is now more information about
1758 building with CMake in the NON-UNIX-USE document.
1759
1760
1761 Version 7.3 28-Aug-07
1762 ---------------------
1763
1764 1. In the rejigging of the build system that eventually resulted in 7.1, the
1765 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1766 brackets there is not right, since it causes compilers to look for an
1767 installed pcre.h, not the version that is in the source that is being
1768 compiled (which of course may be different). I have changed it back to:
1769
1770 #include "pcre.h"
1771
1772 I have a vague recollection that the change was concerned with compiling in
1773 different directories, but in the new build system, that is taken care of
1774 by the VPATH setting the Makefile.
1775
1776 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1777 when the subject happened to end in the byte 0x85 (e.g. if the last
1778 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1779 characters but of course it shouldn't be taken as a newline when it is part
1780 of another character. The bug was that, for an unlimited repeat of . in
1781 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1782 characters when looking for a newline.
1783
1784 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1785
1786 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1787 in debug output.
1788
1789 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1790 long printing in the pcrecpp unittest when running under MinGW.
1791
1792 6. ESC_K was left out of the EBCDIC table.
1793
1794 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1795 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1796 limit also applies to "virtual nesting" when a pattern is recursive, and in
1797 this case 1000 isn't so big. I have been able to remove this limit at the
1798 expense of backing off one optimization in certain circumstances. Normally,
1799 when pcre_exec() would call its internal match() function recursively and
1800 immediately return the result unconditionally, it uses a "tail recursion"
1801 feature to save stack. However, when a subpattern that can match an empty
1802 string has an unlimited repetition quantifier, it no longer makes this
1803 optimization. That gives it a stack frame in which to save the data for
1804 checking that an empty string has been matched. Previously this was taken
1805 from the 1000-entry workspace that had been reserved. So now there is no
1806 explicit limit, but more stack is used.
1807
1808 8. Applied Daniel's patches to solve problems with the import/export magic
1809 syntax that is required for Windows, and which was going wrong for the
1810 pcreposix and pcrecpp parts of the library. These were overlooked when this
1811 problem was solved for the main library.
1812
1813 9. There were some crude static tests to avoid integer overflow when computing
1814 the size of patterns that contain repeated groups with explicit upper
1815 limits. As the maximum quantifier is 65535, the maximum group length was
1816 set at 30,000 so that the product of these two numbers did not overflow a
1817 32-bit integer. However, it turns out that people want to use groups that
1818 are longer than 30,000 bytes (though not repeat them that many times).
1819 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1820 made it possible to implement the integer overflow checks in a much more
1821 dynamic way, which I have now done. The artificial limitation on group
1822 length has been removed - we now have only the limit on the total length of
1823 the compiled pattern, which depends on the LINK_SIZE setting.
1824
1825 10. Fixed a bug in the documentation for get/copy named substring when
1826 duplicate names are permitted. If none of the named substrings are set, the
1827 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1828 empty string.
1829
1830 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1831 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1832 because the ] is interpreted as the first data character and the
1833 terminating ] is not found. PCRE has been made compatible with Perl in this
1834 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1835 cause memory overwriting.
1836
1837 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1838 string has been matched (to stop an infinite loop). It was not recognizing
1839 a conditional subpattern that could match an empty string if that
1840 subpattern was within another subpattern. For example, it looped when
1841 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1842 condition was not nested. This bug has been fixed.
1843
1844 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1845 past the start of the subject in the presence of bytes with the top bit
1846 set, for example "\x8aBCD".
1847
1848 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1849 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1850
1851 14. Optimized (?!) to (*FAIL).
1852
1853 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1854 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1855 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1856 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1857 does: it's just the validity check that is more restrictive.
1858
1859 16. Inserted checks for integer overflows during escape sequence (backslash)
1860 processing, and also fixed erroneous offset values for syntax errors during
1861 backslash processing.
1862
1863 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1864 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1865
1866 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1867 caused an overrun.
1868
1869 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1870 something other than just ASCII characters) inside a group that had an
1871 unlimited repeat caused a loop at compile time (while checking to see
1872 whether the group could match an empty string).
1873
1874 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1875 [\P{Any}] did so. (Error in the code for printing property names.)
1876
1877 21. An orphan \E inside a character class could cause a crash.
1878
1879 22. A repeated capturing bracket such as (A)? could cause a wild memory
1880 reference during compilation.
1881
1882 23. There are several functions in pcre_compile() that scan along a compiled
1883 expression for various reasons (e.g. to see if it's fixed length for look
1884 behind). There were bugs in these functions when a repeated \p or \P was
1885 present in the pattern. These operators have additional parameters compared
1886 with \d, etc, and these were not being taken into account when moving along
1887 the compiled data. Specifically:
1888
1889 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1890 length.
1891
1892 (b) An item such as \pL+ within a repeated group could cause crashes or
1893 loops.
1894
1895 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1896 "reference to non-existent subpattern" error.
1897
1898 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1899
1900 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1901 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1902
1903 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1904
1905 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1906 character were causing crashes (broken optimization).
1907
1908 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1909 \p or \P) caused a compile-time loop.
1910
1911 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1912 break. For example, the unstudied pattern [\r\n]A does not match the string
1913 "\r\nA" because change 7.0/46 below moves the current point on by two
1914 characters after failing to match at the start. However, the pattern \nA
1915 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1916 the same is true. There doesn't seem any very clean way out of this, but
1917 what I have chosen to do makes the common cases work: PCRE now takes note
1918 of whether there can be an explicit match for \r or \n anywhere in the
1919 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1920 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1921 pattern has explicit CR or LF references.
1922
1923 29. Added (*CR) etc for changing newline setting at start of pattern.
1924
1925
1926 Version 7.2 19-Jun-07
1927 ---------------------
1928
1929 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1930 which is apparently normally available under Windows.
1931
1932 2. Re-jig the pcregrep tests with different newline settings in an attempt
1933 to make them independent of the local environment's newline setting.
1934
1935 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1936
1937 4. Some of the "internals" tests were previously cut out when the link size
1938 was not 2, because the output contained actual offsets. The recent new
1939 "Z" feature of pcretest means that these can be cut out, making the tests
1940 usable with all link sizes.
1941
1942 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1943 stack recursion. This gives a massive performance boost under BSD, but just
1944 a small improvement under Linux. However, it saves one field in the frame
1945 in all cases.
1946
1947 6. Added more features from the forthcoming Perl 5.10:
1948
1949 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1950 recursion call. It refers to the nth most recently opened parentheses.
1951
1952 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1953 to be opened parentheses.
1954
1955 (c) Conditions that refer to capturing parentheses can be specified
1956 relatively, for example, (?(-2)... or (?(+3)...
1957
1958 (d) \K resets the start of the current match so that everything before
1959 is not part of it.
1960
1961 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1962
1963 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1964 reference syntax.
1965
1966 (g) (?| introduces a group in which the numbering of parentheses in each
1967 alternative starts with the same number.
1968
1969 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1970
1971 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1972 PCRE_INFO_JCHANGED.
1973
1974 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1975 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1976 for detecting groups that can match an empty string.
1977
1978 9. A pattern with a very large number of alternatives (more than several
1979 hundred) was running out of internal workspace during the pre-compile
1980 phase, where pcre_compile() figures out how much memory will be needed. A
1981 bit of new cunning has reduced the workspace needed for groups with
1982 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1983 workspace instead of running out of the 4096 that are available.
1984
1985 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1986
1987 11. Applied patch from Google to remove an optimization that didn't quite work.
1988 The report of the bug said:
1989
1990 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1991 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1992 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1993
1994 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1995 it matched the wrong number of bytes.
1996
1997
1998 Version 7.1 24-Apr-07
1999 ---------------------
2000
2001 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2002 that is more "standard", making use of automake and other Autotools. There
2003 is some re-arrangement of the files and adjustment of comments consequent
2004 on this.
2005
2006 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2007 for recursive directory scanning broke on some systems because the files
2008 are not scanned in any specific order and on different systems the order
2009 was different. A call to "sort" has been inserted into RunGrepTest for the
2010 approprate test as a short-term fix. In the longer term there may be an
2011 alternative.
2012
2013 3. I had an email from Eric Raymond about problems translating some of PCRE's
2014 man pages to HTML (despite the fact that I distribute HTML pages, some
2015 people do their own conversions for various reasons). The problems
2016 concerned the use of low-level troff macros .br and .in. I have therefore
2017 removed all such uses from the man pages (some were redundant, some could
2018 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2019 HTML has been updated to handle .nf/.fi and to complain if it encounters
2020 .br or .in.
2021
2022 4. Updated comments in configure.ac that get placed in config.h.in and also
2023 arranged for config.h to be included in the distribution, with the name
2024 config.h.generic, for the benefit of those who have to compile without
2025 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2026
2027 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2028 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2029 makevp.bat for the current PCRE, using the additional files
2030 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2031
2032 6. A Windows user reported a minor discrepancy with test 2, which turned out
2033 to be caused by a trailing space on an input line that had got lost in his
2034 copy. The trailing space was an accident, so I've just removed it.
2035
2036 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2037 that is needed.
2038
2039 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2040 as "const" (a) because they are and (b) because it helps the PHP
2041 maintainers who have recently made a script to detect big data structures
2042 in the php code that should be moved to the .rodata section. I remembered
2043 to update Builducptable as well, so it won't revert if ucptable.h is ever
2044 re-created.
2045
2046 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2047 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2048 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2049 support is not required. This saves 1.5-2K of code, which is important in
2050 some applications.
2051
2052 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2053 so as not to refer to the tables, even though these functions will never be
2054 called when UTF-8 support is disabled. Otherwise there are problems with a
2055 shared library.
2056
2057 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2058
2059 (a) It was defining its arguments as char * instead of void *.
2060
2061 (b) It was assuming that all moves were upwards in memory; this was true
2062 a long time ago when I wrote it, but is no longer the case.
2063
2064 The emulated memove() is provided for those environments that have neither
2065 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2066 is clearly not the case, as these two bugs were recently reported.
2067
2068 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2069 and Detrail to create the HTML documentation, the .txt form of the man
2070 pages, and it removes trailing spaces from listed files. It also creates
2071 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2072 case, it wraps all the #defines with #ifndefs. This script should be run
2073 before "make dist".
2074
2075 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2076 with Unicode property support.
2077
2078 (a) For a maximizing quantifier, if the two different cases of the
2079 character were of different lengths in their UTF-8 codings (there are
2080 some cases like this - I found 11), and the matching function had to
2081 back up over a mixture of the two cases, it incorrectly assumed they
2082 were both the same length.
2083
2084 (b) When PCRE was configured to use the heap rather than the stack for
2085 recursion during matching, it was not correctly preserving the data for
2086 the other case of a UTF-8 character when checking ahead for a match
2087 while processing a minimizing repeat. If the check also involved
2088 matching a wide character, but failed, corruption could cause an
2089 erroneous result when trying to check for a repeat of the original
2090 character.
2091
2092 13. Some tidying changes to the testing mechanism:
2093
2094 (a) The RunTest script now detects the internal link size and whether there
2095 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2096 values substituted by "configure". (The RunGrepTest script already did
2097 this for UTF-8.) The configure.ac script no longer substitutes the
2098 relevant variables.
2099
2100 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2101 with length and offset values. This means that the output is different
2102 for different internal link sizes. Test 2 is skipped for link sizes
2103 other than 2 because of this, bypassing the problem. Unfortunately,
2104 there was also a test in test 3 (the locale tests) that used /B and
2105 failed for link sizes other than 2. Rather than cut the whole test out,
2106 I have added a new /Z option to pcretest that replaces the length and
2107 offset values with spaces. This is now used to make test 3 independent
2108 of link size. (Test 2 will be tidied up later.)
2109
2110 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2111 segmentation fault instead of returning the appropriate error message.
2112
2113 15. In multiline mode when the newline sequence was set to "any", the pattern
2114 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2115 This doesn't seem right; it now treats the CRLF combination as the line
2116 ending, and so does not match in that case. It's only a pattern such as ^$
2117 that would hit this one: something like ^ABC$ would have failed after \r
2118 and then tried again after \r\n.
2119
2120 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2121 in an attempt to make files that differ only in their line terminators
2122 compare equal. This works on Linux.
2123
2124 17. Under certain error circumstances pcregrep might try to free random memory
2125 as it exited. This is now fixed, thanks to valgrind.
2126
2127 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2128 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2129 was because its rules for how to advance for /g after matching an empty
2130 string at the end of a line did not allow for this case. They now check for
2131 it specially.
2132
2133 20. pcretest is supposed to handle patterns and data of any length, by
2134 extending its buffers when necessary. It was getting this wrong when the
2135 buffer for a data line had to be extended.
2136
2137 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2138 CRLF as a newline sequence.
2139
2140 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2141 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2142 I have nevertheless tidied it up.
2143
2144 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2145
2146 24. Added a man page for pcre-config.
2147
2148
2149 Version 7.0 19-Dec-06
2150 ---------------------
2151
2152 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2153 moving to gcc 4.1.1.
2154
2155 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2156 sys/time.h, which is documented as needed for this function. It doesn't
2157 seem to matter on Linux, but it showed up on some releases of OS X.
2158
2159 3. It seems that there are systems where bytes whose values are greater than
2160 127 match isprint() in the "C" locale. The "C" locale should be the
2161 default when a C program starts up. In most systems, only ASCII printing
2162 characters match isprint(). This difference caused the output from pcretest
2163 to vary, making some of the tests fail. I have changed pcretest so that:
2164
2165 (a) When it is outputting text in the compiled version of a pattern, bytes
2166 other than 32-126 are always shown as hex escapes.
2167
2168 (b) When it is outputting text that is a matched part of a subject string,
2169 it does the same, unless a different locale has been set for the match
2170 (using the /L modifier). In this case, it uses isprint() to decide.
2171
2172 4. Fixed a major bug that caused incorrect computation of the amount of memory
2173 required for a compiled pattern when options that changed within the
2174 pattern affected the logic of the preliminary scan that determines the
2175 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2176 that the computed length was too small. The symptoms of this bug were
2177 either the PCRE error "internal error: code overflow" from pcre_compile(),
2178 or a glibc crash with a message such as "pcretest: free(): invalid next
2179 size (fast)". Examples of patterns that provoked this bug (shown in
2180 pcretest format) are:
2181
2182 /(?-x: )/x
2183 /(?x)(?-x: \s*#\s*)/
2184 /((?i)[\x{c0}])/8
2185 /(?i:[\x{c0}])/8
2186
2187 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2188 is now done differently.
2189
2190 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2191 wrapper classes; (b) implement a new function in the C++ scanner that is
2192 more efficient than the old way of doing things because it avoids levels of
2193 recursion in the regex matching; (c) add a paragraph to the documentation
2194 for the FullMatch() function.
2195
2196 6. The escape sequence \n was being treated as whatever was defined as
2197 "newline". Not only was this contrary to the documentation, which states
2198 that \n is character 10 (hex 0A), but it also went horribly wrong when
2199 "newline" was defined as CRLF. This has been fixed.
2200
2201 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2202 was being set to -1 for the "end of line" case (supposedly a value that no
2203 character can have). Though this value is never used (the check for end of
2204 line is "zero bytes in current character"), it caused compiler complaints.
2205 I've changed it to 0xffffffff.
2206
2207 8. In pcre_version.c, the version string was being built by a sequence of
2208 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2209 string (as it is for production releases) called a macro with an empty
2210 argument. The C standard says the result of this is undefined. The gcc
2211 compiler treats it as an empty string (which was what was wanted) but it is
2212 reported that Visual C gives an error. The source has been hacked around to
2213 avoid this problem.
2214
2215 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2216 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2217 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2218 of them did).
2219
2220 10. Originally, pcretest opened its input and output without "b"; then I was
2221 told that "b" was needed in some environments, so it was added for release
2222 5.0 to both the input and output. (It makes no difference on Unix-like
2223 systems.) Later I was told that it is wrong for the input on Windows. I've
2224 now abstracted the modes into two macros, to make it easier to fiddle with
2225 them, and removed "b" from the input mode under Windows.
2226
2227 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2228
2229 12. Added -help and --help to pcretest as an official way of being reminded
2230 of the options.
2231
2232 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2233 and pcrecpp.cc because they annoy compilers at high warning levels.
2234
2235 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2236
2237 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2238 scripts are not C programs :-) and which was not noticed because it works
2239 on Linux.
2240
2241 16. pcretest is supposed to handle any length of pattern and data line (as one
2242 line or as a continued sequence of lines) by extending its input buffer if
2243 necessary. This feature was broken for very long pattern lines, leading to
2244 a string of junk being passed to pcre_compile() if the pattern was longer
2245 than about 50K.
2246
2247 17. I have done a major re-factoring of the way pcre_compile() computes the
2248 amount of memory needed for a compiled pattern. Previously, there was code
2249 that made a preliminary scan of the pattern in order to do this. That was
2250 OK when PCRE was new, but as the facilities have expanded, it has become
2251 harder and harder to keep it in step with the real compile phase, and there
2252 have been a number of bugs (see for example, 4 above). I have now found a
2253 cunning way of running the real compile function in a "fake" mode that
2254 enables it to compute how much memory it would need, while actually only
2255 ever using a few hundred bytes of working memory and without too many
2256 tests of the mode. This should make future maintenance and development
2257 easier. A side effect of this work is that the limit of 200 on the nesting
2258 depth of parentheses has been removed (though this was never a serious
2259 limitation, I suspect). However, there is a downside: pcre_compile() now
2260 runs more slowly than before (30% or more, depending on the pattern). I
2261 hope this isn't a big issue. There is no effect on runtime performance.
2262
2263 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2264 newline (only possible for the last line of a file) and it was a
2265 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2266
2267 19. Added additional timing features to pcretest. (1) The -tm option now times
2268 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2269 separate command line item, by a number that specifies the number of
2270 repeats to use when timing. The default is 50000; this gives better
2271 precision, but takes uncomfortably long for very large patterns.
2272
2273 20. Extended pcre_study() to be more clever in cases where a branch of a
2274 subpattern has no definite first character. For example, (a*|b*)[cd] would
2275 previously give no result from pcre_study(). Now it recognizes that the
2276 first character must be a, b, c, or d.
2277
2278 21. There was an incorrect error "recursive call could loop indefinitely" if
2279 a subpattern (or the entire pattern) that was being tested for matching an
2280 empty string contained only one non-empty item after a nested subpattern.
2281 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2282 incorrectly, because the \d was being skipped in the check.
2283
2284 22. The pcretest program now has a new pattern option /B and a command line
2285 option -b, which is equivalent to adding /B to every pattern. This causes
2286 it to show the compiled bytecode, without the additional information that
2287 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2288 is the same as /B/I).
2289
2290 23. A new optimization is now able automatically to treat some sequences such
2291 as a*b as a*+b. More specifically, if something simple (such as a character
2292 or a simple class like \d) has an unlimited quantifier, and is followed by
2293 something that cannot possibly match the quantified thing, the quantifier
2294 is automatically "possessified".
2295
2296 24. A recursive reference to a subpattern whose number was greater than 39
2297 went wrong under certain circumstances in UTF-8 mode. This bug could also
2298 have affected the operation of pcre_study().
2299
2300 25. Realized that a little bit of performance could be had by replacing
2301 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2302
2303 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2304
2305 27. Possessive quantifiers such as a++ were previously implemented by turning
2306 them into atomic groups such as ($>a+). Now they have their own opcodes,
2307 which improves performance. This includes the automatically created ones
2308 from 23 above.
2309
2310 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2311 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2312 the first matched character to be a colon. This applied both to named and
2313 numbered groups.
2314
2315 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2316
2317 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2318 building PCRE on Windows easier, so I have included it in the distribution.
2319
2320 31. There is now a check in pcretest against a ridiculously large number being
2321 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2322 loop, the loop is abandoned.
2323
2324 32. Forward references to subpatterns in conditions such as (?(2)...) where
2325 subpattern 2 is defined later cause pcre_compile() to search forwards in
2326 the pattern for the relevant set of parentheses. This search went wrong
2327 when there were unescaped parentheses in a character class, parentheses
2328 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2329
2330 33. "Subroutine" calls and backreferences were previously restricted to
2331 referencing subpatterns earlier in the regex. This restriction has now
2332 been removed.
2333
2334 34. Added a number of extra features that are going to be in Perl 5.10. On the
2335 whole, these are just syntactic alternatives for features that PCRE had
2336 previously implemented using the Python syntax or my own invention. The
2337 other formats are all retained for compatibility.
2338
2339 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2340 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2341 also .NET compatible.
2342
2343 (b) A recursion or subroutine call to a named group can now be defined as
2344 (?&name) as well as (?P>name).
2345
2346 (c) A backreference to a named group can now be defined as \k<name> or
2347 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2348 5.10, are also .NET compatible.
2349
2350 (d) A conditional reference to a named group can now use the syntax
2351 (?(<name>) or (?('name') as well as (?(name).
2352
2353 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2354 groups (named and numbered) that are never evaluated inline, but can be
2355 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2356 is always false. There may be only one alternative in such a group.
2357
2358 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2359 as the simple (?(R). The condition is true only if the most recent
2360 recursion is that of the given number or name. It does not search out
2361 through the entire recursion stack.
2362
2363 (g) The escape \gN or \g{N} has been added, where N is a positive or
2364 negative number, specifying an absolute or relative reference.
2365
2366 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2367 some "unreachable code" warnings.
2368
2369 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2370 things, this adds five new scripts.
2371
2372 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2373 There were also incompatibilities regarding the handling of \Q..\E inside
2374 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2375 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2376
2377 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2378 matches an empty string, and forcibly breaks the loop. There were bugs in
2379 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2380 against aaaa the result was just "a" rather than "aaaa", for example. Two
2381 separate and independent bugs (that affected different cases) have been
2382 fixed.
2383
2384 39. Refactored the code to abolish the use of different opcodes for small
2385 capturing bracket numbers. This is a tidy that I avoided doing when I
2386 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2387 The new approach is not only tidier, it makes it possible to reduce the
2388 memory needed to fix the previous bug (38).
2389
2390 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2391 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2392 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2393 mode.
2394
2395 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2396 report.
2397
2398 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2399 copy construction and assignment in the C++ wrapper.
2400
2401 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2402 couple of bugs that could have given wrong results in the "--newline=crlf"
2403 case.
2404
2405 44. Added a number of casts and did some reorganization of signed/unsigned int
2406 variables following suggestions from Dair Grant. Also renamed the variable
2407 "this" as "item" because it is a C++ keyword.
2408
2409 45. Arranged for dftables to add
2410
2411 #include "pcre_internal.h"
2412
2413 to pcre_chartables.c because without it, gcc 4.x may remove the array
2414 definition from the final binary if PCRE is built into a static library and
2415 dead code stripping is activated.
2416
2417 46. For an unanchored pattern, if a match attempt fails at the start of a
2418 newline sequence, and the newline setting is CRLF or ANY, and the next two
2419 characters are CRLF, advance by two characters instead of one.
2420
2421
2422 Version 6.7 04-Jul-06
2423 ---------------------
2424
2425 1. In order to handle tests when input lines are enormously long, pcretest has
2426 been re-factored so that it automatically extends its buffers when
2427 necessary. The code is crude, but this _is_ just a test program. The
2428 default size has been increased from 32K to 50K.
2429
2430 2. The code in pcre_study() was using the value of the re argument before
2431 testing it for NULL. (Of course, in any sensible call of the function, it
2432 won't be NULL.)
2433
2434 3. The memmove() emulation function in pcre_internal.h, which is used on
2435 systems that lack both memmove() and bcopy() - that is, hardly ever -
2436 was missing a "static" storage class specifier.
2437
2438 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2439 containing an extended class (one that cannot be represented by a bitmap
2440 because it contains high-valued characters or Unicode property items, e.g.
2441 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2442 pattern, but PCRE should not loop if you do not (it no longer does).
2443 [Detail: two cases were found: (a) a repeated subpattern containing an
2444 extended class; (b) a recursive reference to a subpattern that followed a
2445 previous extended class. It wasn't skipping over the extended class
2446 correctly when UTF-8 mode was not set.]
2447
2448 5. A negated single-character class was not being recognized as fixed-length
2449 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2450 compile error "lookbehind assertion is not fixed length".
2451
2452 6. The RunPerlTest auxiliary script was showing an unexpected difference
2453 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2454 write a Perl script that can interpret lines of an input file either as
2455 byte characters or as UTF-8, which is what "perltest" was being required to
2456 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2457 can't do is switch easily at run time between having the "use utf8;" pragma
2458 or not. In the end, I fudged it by using the RunPerlTest script to insert
2459 "use utf8;" explicitly for the UTF-8 tests.
2460
2461 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2462 the end of the subject string, contrary to the documentation and to what
2463 Perl does. This was true of both matching functions. Now it matches only at
2464 the start of the subject and immediately after *internal* newlines.
2465
2466 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2467 a pointer to an int instead of a pointer to an unsigned long int. This
2468 caused problems on 64-bit systems.
2469
2470 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2471 instance of the 'standard' template library not being so standard".
2472
2473 10. There was no check on the number of named subpatterns nor the maximum
2474 length of a subpattern name. The product of these values is used to compute
2475 the size of the memory block for a compiled pattern. By supplying a very
2476 long subpattern name and a large number of named subpatterns, the size
2477 computation could be caused to overflow. This is now prevented by limiting
2478 the length of names to 32 characters, and the number of named subpatterns
2479 to 10,000.
2480
2481 11. Subpatterns that are repeated with specific counts have to be replicated in
2482 the compiled pattern. The size of memory for this was computed from the
2483 length of the subpattern and the repeat count. The latter is limited to
2484 65535, but there was no limit on the former, meaning that integer overflow
2485 could in principle occur. The compiled length of a repeated subpattern is
2486 now limited to 30,000 bytes in order to prevent this.
2487
2488 12. Added the optional facility to have named substrings with the same name.
2489
2490 13. Added the ability to use a named substring as a condition, using the
2491 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2492 are numbers (not recommended). Forward references are permitted.
2493
2494 14. Added forward references in named backreferences (if you see what I mean).
2495
2496 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2497 pattern could run off the end of the subject. For example, the pattern
2498 "(?s)(.{1,5})"8 did this with the subject "ab".
2499
2500 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2501 PCRE_CASELESS was set when matching characters that were quantified with ?
2502 or *.
2503
2504 17. A character class other than a single negated character that had a minimum
2505 but no maximum quantifier - for example [ab]{6,} - was not handled
2506 correctly by pce_dfa_exec(). It would match only one character.
2507
2508 18. A valid (though odd) pattern that looked like a POSIX character
2509 class but used an invalid character after [ (for example [[,abc,]]) caused
2510 pcre_compile() to give the error "Failed: internal error: code overflow" or
2511 in some cases to crash with a glibc free() error. This could even happen if
2512 the pattern terminated after [[ but there just happened to be a sequence of
2513 letters, a binary zero, and a closing ] in the memory that followed.
2514
2515 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2516 over the years. Originally (before any Unicode support), just the bottom 8
2517 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2518 output from "man perlunicode" includes this:
2519
2520 The regular expression compiler produces polymorphic opcodes. That
2521 is, the pattern adapts to the data and automatically switches to
2522 the Unicode character scheme when presented with Unicode data--or
2523 instead uses a traditional byte scheme when presented with byte
2524 data.
2525
2526 Sadly, a wide octal escape does not cause a switch, and in a string with
2527 no other multibyte characters, these octal escapes are treated as before.
2528 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2529 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2530 Unicode string.
2531
2532 I have not perpetrated such confusion in PCRE. Up till now, it took just
2533 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2534 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2535 translate to the appropriate multibyte character.
2536
2537 29. Applied some refactoring to reduce the number of warnings from Microsoft
2538 and Borland compilers. This has included removing the fudge introduced
2539 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2540 a warning about an unused variable.
2541
2542 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2543 characters since release 4.0, because Perl (from release 5.004) does not.
2544 [Or at least, is documented not to: some releases seem to be in conflict
2545 with the documentation.] However, when a pattern was studied with
2546 pcre_study() and all its branches started with \s, PCRE still included VT
2547 as a possible starting character. Of course, this did no harm; it just
2548 caused an unnecessary match attempt.
2549
2550 22. Removed a now-redundant internal flag bit that recorded the fact that case
2551 dependency changed within the pattern. This was once needed for "required
2552 byte" processing, but is no longer used. This recovers a now-scarce options
2553 bit. Also moved the least significant internal flag bit to the most-
2554 significant bit of the word, which was not previously used (hangover from
2555 the days when it was an int rather than a uint) to free up another bit for
2556 the future.
2557
2558 23. Added support for CRLF line endings as well as CR and LF. As well as the
2559 default being selectable at build time, it can now be changed at runtime
2560 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2561 specify that it is scanning data with non-default line endings.
2562
2563 24. Changed the definition of CXXLINK to make it agree with the definition of
2564 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2565
2566 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2567 recursions. This makes a big different to stack usage for some patterns.
2568
2569 26. If a subpattern containing a named recursion or subroutine reference such
2570 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2571 the space required for the compiled pattern went wrong and gave too small a
2572 value. Depending on the environment, this could lead to "Failed: internal
2573 error: code overflow at offset 49" or "glibc detected double free or
2574 corruption" errors.
2575
2576 27. Applied patches from Google (a) to support the new newline modes and (b) to
2577 advance over multibyte UTF-8 characters in GlobalReplace.
2578
2579 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2580 difference for some implementation of PCRE in some Windows version.
2581
2582 29. Added some extra testing facilities to pcretest:
2583
2584 \q<number> in a data line sets the "match limit" value
2585 \Q<number> in a data line sets the "match recursion limt" value
2586 -S <number> sets the stack size, where <number> is in megabytes
2587
2588 The -S option isn't available for Windows.
2589
2590
2591 Version 6.6 06-Feb-06
2592 ---------------------
2593
2594 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2595 in pcreposix.h. I have copied the definition from pcre.h.
2596
2597 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2598 because pcre.h is no longer a built file.
2599
2600 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2601 not normally included in the compiled code.
2602
2603
2604 Version 6.5 01-Feb-06
2605 ---------------------
2606
2607 1. When using the partial match feature with pcre_dfa_exec(), it was not
2608 anchoring the second and subsequent partial matches at the new starting
2609 point. This could lead to incorrect results. For example, with the pattern
2610 /1234/, partially matching against "123" and then "a4" gave a match.
2611
2612 2. Changes to pcregrep:
2613
2614 (a) All non-match returns from pcre_exec() were being treated as failures
2615 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2616 error message is output. Some extra information is given for the
2617 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2618 probably the only errors that are likely to be caused by users (by
2619 specifying a regex that has nested indefinite repeats, for instance).
2620 If there are more than 20 of these errors, pcregrep is abandoned.
2621
2622 (b) A binary zero was treated as data while matching, but terminated the
2623 output line if it was written out. This has been fixed: binary zeroes
2624 are now no different to any other data bytes.
2625
2626 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2627 used to set a locale for matching. The --locale=xxxx long option has
2628 been added (no short equivalent) to specify a locale explicitly on the
2629 pcregrep command, overriding the environment variables.
2630
2631 (d) When -B was used with -n, some line numbers in the output were one less
2632 than they should have been.
2633
2634 (e) Added the -o (--only-matching) option.
2635
2636 (f) If -A or -C was used with -c (count only), some lines of context were
2637 accidentally printed for the final match.
2638
2639 (g) Added the -H (--with-filename) option.
2640
2641 (h) The combination of options -rh failed to suppress file names for files
2642 that were found from directory arguments.
2643
2644 (i) Added the -D (--devices) and -d (--directories) options.
2645
2646 (j) Added the -F (--fixed-strings) option.
2647
2648 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2649
2650 (l) Added the --colo(u)r option.
2651
2652 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2653 is not present by default.
2654
2655 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2656 items such as (?R) or (?1), when the recursion could match a number of
2657 alternatives. If it matched one of the alternatives, but subsequently,
2658 outside the recursion, there was a failure, the code tried to back up into
2659 the recursion. However, because of the way PCRE is implemented, this is not
2660 possible, and the result was an incorrect result from the match.
2661
2662 In order to prevent this happening, the specification of recursion has
2663 been changed so that all such subpatterns are automatically treated as
2664 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2665
2666 4. I had overlooked the fact that, in some locales, there are characters for
2667 which isalpha() is true but neither isupper() nor islower() are true. In
2668 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2669 and ordfeminine) are like this. This affected the treatment of \w and \W
2670 when they appeared in character classes, but not when they appeared outside
2671 a character class. The bit map for "word" characters is now created
2672 separately from the results of isalnum() instead of just taking it from the
2673 upper, lower, and digit maps. (Plus the underscore character, of course.)
2674
2675 5. The above bug also affected the handling of POSIX character classes such as
2676 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2677 permanent tables. Instead, the bit maps for such a class were previously
2678 created as the appropriate unions of the upper, lower, and digit bitmaps.
2679 Now they are created by subtraction from the [[:word:]] class, which has
2680 its own bitmap.
2681
2682 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2683 It is created by subtracting the vertical space characters (\x09, \x0a,
2684 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2685 subtraction was done in the overall bitmap for a character class, meaning
2686 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2687 be recognized. This bug has been fixed.
2688
2689 7. Patches from the folks at Google:
2690
2691 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2692 real life, but is still worth protecting against".
2693
2694 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2695 regular expressions".
2696
2697 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2698 have it.
2699
2700 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2701 "configure" and the latter not, in order to fix a problem somebody had
2702 with compiling the Arg class on HP-UX.
2703
2704 (e) Improve the error-handling of the C++ wrapper a little bit.
2705
2706 (f) New tests for checking recursion limiting.
2707
2708 8. The pcre_memmove() function, which is used only if the environment does not
2709 have a standard memmove() function (and is therefore rarely compiled),
2710 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2711 returning a result (though PCRE never actually uses the result).
2712
2713 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2714 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2715 returned instead of calling malloc() with an overflowing number that would
2716 most likely cause subsequent chaos.
2717
2718 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2719
2720 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2721 with this option is matched, the nmatch and pmatch options of regexec() are
2722 ignored.
2723
2724 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2725 provided in case anyone wants to the the POSIX interface with UTF-8
2726 strings.
2727
2728 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2729 C++ linking (needed for some HP-UX environments).
2730
2731 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2732 (unused parameter) and in the pcre_printint() function (omitted "default"
2733 switch label when the default is to do nothing).
2734
2735 15. Added some code to make it possible, when PCRE is compiled as a C++
2736 library, to replace subject pointers for pcre_exec() with a smart pointer
2737 class, thus making it possible to process discontinuous strings.
2738
2739 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2740 much the same function. They were added by different people who were trying
2741 to make PCRE easy to compile on non-Unix systems. It has been suggested
2742 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2743 for compiling on Windows systems. I have therefore replaced it with
2744 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2745 defaults to "extern" for C or "extern C" for C++, which works fine on
2746 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2747 SCOPE with something explicit in config.h. In addition:
2748
2749 (a) pcreposix.h still had just "extern" instead of either of these macros;
2750 I have replaced it with PCRE_DATA_SCOPE.
2751
2752 (b) Functions such as _pcre_xclass(), which are internal to the library,
2753 but external in the C sense, all had PCRE_EXPORT in their definitions.
2754 This is apparently wrong for the Windows case, so I have removed it.
2755 (It makes no difference on Unix-like systems.)
2756
2757 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2758 of recursive calls to match(). This is different to MATCH_LIMIT because
2759 that limits the total number of calls to match(), not all of which increase
2760 the depth of recursion. Limiting the recursion depth limits the amount of
2761 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2762 when PCRE is compiled, and changed at run time. A patch from Google adds
2763 this functionality to the C++ interface.
2764
2765 18. Changes to the handling of Unicode character properties:
2766
2767 (a) Updated the table to Unicode 4.1.0.
2768
2769 (b) Recognize characters that are not in the table as "Cn" (undefined).
2770
2771 (c) I revised the way the table is implemented to a much improved format
2772 which includes recognition of ranges. It now supports the ranges that
2773 are defined in UnicodeData.txt, and it also amalgamates other
2774 characters into ranges. This has reduced the number of entries in the
2775 table from around 16,000 to around 3,000, thus reducing its size
2776 considerably. I realized I did not need to use a tree structure after
2777 all - a binary chop search is just as efficient. Having reduced the
2778 number of entries, I extended their size from 6 bytes to 8 bytes to
2779 allow for more data.
2780
2781 (d) Added support for Unicode script names via properties such as \p{Han}.
2782
2783 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2784 matching that character.
2785
2786 20. When matching a repeated Unicode property with a minimum greater than zero,
2787 (for example \pL{2,}), PCRE could look past the end of the subject if it
2788 reached it while seeking the minimum number of characters. This could
2789 happen only if some of the characters were more than one byte long, because
2790 there is a check for at least the minimum number of bytes.
2791
2792 21. Refactored the implementation of \p and \P so as to be more general, to
2793 allow for more different types of property in future. This has changed the
2794 compiled form incompatibly. Anybody with saved compiled patterns that use
2795 \p or \P will have to recompile them.
2796
2797 22. Added "Any" and "L&" to the supported property types.
2798
2799 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2800 but give a compile time error if the value is greater than 0xff.
2801
2802 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2803 accidentally not being installed or uninstalled.
2804
2805 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2806 made were to insert the current release number. This seemed silly, because
2807 it made things harder for people building PCRE on systems that don't run
2808 "configure". I have turned pcre.h into a distributed file, no longer built
2809 by "configure", with the version identification directly included. There is
2810 no longer a pcre.h.in file.
2811
2812 However, this change necessitated a change to the pcre-config script as
2813 well. It is built from pcre-config.in, and one of the substitutions was the
2814 release number. I have updated configure.ac so that ./configure now finds
2815 the release number by grepping pcre.h.
2816
2817 26. Added the ability to run the tests under valgrind.
2818
2819
2820 Version 6.4 05-Sep-05
2821 ---------------------
2822
2823 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2824 "--" to be printed when multiple files were scanned, even when none of the
2825 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2826 consider it to be a bug, and have restored the previous behaviour.
2827
2828 2. A couple of code tidies to get rid of compiler warnings.
2829
2830 3. The pcretest program used to cheat by referring to symbols in the library
2831 whose names begin with _pcre_. These are internal symbols that are not
2832 really supposed to be visible externally, and in some environments it is
2833 possible to suppress them. The cheating is now confined to including
2834 certain files from the library's source, which is a bit cleaner.
2835
2836 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2837 file's purpose clearer.
2838
2839 5. Reorganized pcre_ucp_findchar().
2840
2841
2842 Version 6.3 15-Aug-05
2843 ---------------------
2844
2845 1. The file libpcre.pc.in did not have general read permission in the tarball.
2846
2847 2. There were some problems when building without C++ support:
2848
2849 (a) If C++ support was not built, "make install" and "make test" still
2850 tried to test it.
2851
2852 (b) There were problems when the value of CXX was explicitly set. Some
2853 changes have been made to try to fix these, and ...
2854
2855 (c) --disable-cpp can now be used to explicitly disable C++ support.
2856
2857 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2858 backslash in a target when C++ was disabled. This confuses some
2859 versions of "make", apparently. Using an intermediate variable solves
2860 this. (Same for CPP_LOBJ.)
2861
2862 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2863 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2864 necessary on certain architectures.
2865
2866 4. Added a setting of -export-symbols-regex to the link command to remove
2867 those symbols that are exported in the C sense, but actually are local
2868 within the library, and not documented. Their names all begin with
2869 "_pcre_". This is not a perfect job, because (a) we have to except some
2870 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2871 available (and never for static libraries). I have made a note to try to
2872 find a way round (a) in the future.
2873
2874
2875 Version 6.2 01-Aug-05
2876 ---------------------
2877
2878 1. There was no test for integer overflow of quantifier values. A construction
2879 such as {1111111111111111} would give undefined results. What is worse, if
2880 a minimum quantifier for a parenthesized subpattern overflowed and became
2881 negative, the calculation of the memory size went wrong. This could have
2882 led to memory overwriting.
2883
2884 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2885
2886 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2887 operating environments where this matters.
2888
2889 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2890 PCRE options from within the C++ wrapper.
2891
2892 5. Named capturing subpatterns were not being correctly counted when a pattern
2893 was compiled. This caused two problems: (a) If there were more than 100
2894 such subpatterns, the calculation of the memory needed for the whole
2895 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2896 back references of the form \12, where the number was greater than 9, were
2897 not recognized as back references, even though there were sufficient
2898 previous subpatterns.
2899
2900 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2901 versions of gcc, e.g. 2.95.4.
2902
2903
2904 Version 6.1 21-Jun-05
2905 ---------------------
2906
2907 1. There was one reference to the variable "posix" in pcretest.c that was not
2908 surrounded by "#if !defined NOPOSIX".
2909
2910 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2911 the cross-check on the old pcre_info() function, for the benefit of the
2912 cut-down version of PCRE that is currently imported into Exim.
2913
2914 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2915 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2916 patterns that start (?i) but I don't think that matters. The use of (?i) is
2917 just an example; this all applies to the other options as well.
2918
2919 4. Since libtool seems to echo the compile commands it is issuing, the output
2920 from "make" can be reduced a bit by putting "@" in front of each libtool
2921 compile command.
2922
2923 5. Patch from the folks at Google for configure.in to be a bit more thorough
2924 in checking for a suitable C++ installation before trying to compile the
2925 C++ stuff. This should fix a reported problem when a compiler was present,
2926 but no suitable headers.
2927
2928 6. The man pages all had just "PCRE" as their title. I have changed them to
2929 be the relevant file name. I have also arranged that these names are
2930 retained in the file doc/pcre.txt, which is a concatenation in text format
2931 of all the man pages except the little individual ones for each function.
2932
2933 7. The NON-UNIX-USE file had not been updated for the different set of source
2934 files that come with release 6. I also added a few comments about the C++
2935 wrapper.
2936
2937
2938 Version 6.0 07-Jun-05
2939 ---------------------
2940
2941 1. Some minor internal re-organization to help with my DFA experiments.
2942
2943 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2944 didn't matter for the library itself when fully configured, but did matter
2945 when compiling without UCP support, or within Exim, where the ucp files are
2946 not imported.
2947
2948 3. Refactoring of the library code to split up the various functions into
2949 different source modules. The addition of the new DFA matching code (see
2950 below) to a single monolithic source would have made it really too
2951 unwieldy, quite apart from causing all the code to be include in a
2952 statically linked application, when only some functions are used. This is
2953 relevant even without the DFA addition now that patterns can be compiled in
2954 one application and matched in another.
2955
2956 The downside of splitting up is that there have to be some external
2957 functions and data tables that are used internally in different modules of
2958 the library but which are not part of the API. These have all had their
2959 names changed to start with "_pcre_" so that they are unlikely to clash
2960 with other external names.
2961
2962 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2963 a different (DFA) algorithm. Although it is slower than the original
2964 function, it does have some advantages for certain types of matching
2965 problem.
2966
2967 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2968 including restarting after a partial match.
2969
2970 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2971 defined when compiling for Windows was sent to me. I have put it into the
2972 code, though I have no means of testing or verifying it.
2973
2974 7. Added the pcre_refcount() auxiliary function.
2975
2976 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2977 match before or at the first newline in the subject string. In pcretest,
2978 the /f option on a pattern can be used to set this.
2979
2980 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2981 would behave wrongly. This has been present in PCRE since release 4.0.
2982
2983 10. A number of changes to the pcregrep command:
2984
2985 (a) Refactored how -x works; insert ^(...)$ instead of setting
2986 PCRE_ANCHORED and checking the length, in preparation for adding
2987 something similar for -w.
2988
2989 (b) Added the -w (match as a word) option.
2990
2991 (c) Refactored the way lines are read and buffered so as to have more
2992 than one at a time available.
2993
2994 (d) Implemented a pcregrep test script.
2995
2996 (e) Added the -M (multiline match) option. This allows patterns to match
2997 over several lines of the subject. The buffering ensures that at least
2998 8K, or the rest of the document (whichever is the shorter) is available
2999 for matching (and similarly the previous 8K for lookbehind assertions).
3000
3001 (f) Changed the --help output so that it now says
3002
3003 -w, --word-regex(p)
3004
3005 instead of two lines, one with "regex" and the other with "regexp"
3006 because that confused at least one person since the short forms are the
3007 same. (This required a bit of code, as the output is generated
3008 automatically from a table. It wasn't just a text change.)
3009
3010 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3011 option but starts with a hyphen. Could be a pattern or a path name
3012 starting with a hyphen, for instance.
3013
3014 (h) "-" can be given as a file name to represent stdin.
3015
3016 (i) When file names are being printed, "(standard input)" is used for
3017 the standard input, for compatibility with GNU grep. Previously
3018 "<stdin>" was used.
3019
3020 (j) The option --label=xxx can be used to supply a name to be used for
3021 stdin when file names are being printed. There is no short form.
3022
3023 (k) Re-factored the options decoding logic because we are going to add
3024 two more options that take data. Such options can now be given in four
3025 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3026
3027 (l) Added the -A, -B, and -C options for requesting that lines of context
3028 around matches be printed.
3029
3030 (m) Added the -L option to print the names of files that do not contain
3031 any matching lines, that is, the complement of -l.
3032
3033 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3034 continue to scan other files.
3035
3036 (o) The -s option was incorrectly implemented. For compatibility with other
3037 greps, it now suppresses the error message for a non-existent or non-
3038 accessible file (but not the return code). There is a new option called
3039 -q that suppresses the output of matching lines, which was what -s was
3040 previously doing.
3041
3042 (p) Added --include and --exclude options to specify files for inclusion
3043 and exclusion when recursing.
3044
3045 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3046 Hopefully, it now does.
3047
3048 12. Missing cast in pcre_study().
3049
3050 13. Added an "uninstall" target to the makefile.
3051
3052 14. Replaced "extern" in the function prototypes in Makefile.in with
3053 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3054 world, but is set differently for Windows.
3055
3056 15. Added a second compiling function called pcre_compile2(). The only
3057 difference is that it has an extra argument, which is a pointer to an
3058 integer error code. When there is a compile-time failure, this is set
3059 non-zero, in addition to the error test pointer being set to point to an
3060 error message. The new argument may be NULL if no error number is required
3061 (but then you may as well call pcre_compile(), which is now just a
3062 wrapper). This facility is provided because some applications need a
3063 numeric error indication, but it has also enabled me to tidy up the way
3064 compile-time errors are handled in the POSIX wrapper.
3065
3066 16. Added VPATH=.libs to the makefile; this should help when building with one
3067 prefix path and installing with another. (Or so I'm told by someone who
3068 knows more about this stuff than I do.)
3069
3070 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3071 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3072 match everything, including newlines. This is not POSIX-compatible, but
3073 somebody wanted the feature. From pcretest it can be activated by using
3074 both the P and the s flags.
3075
3076 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3077
3078 19. libpcre.pc was being incorrectly installed as executable.
3079
3080 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3081 it now also looks for '\r' so that it will work unmodified on Windows.
3082
3083 21. Added Google's contributed C++ wrapper to the distribution.
3084
3085 22. Added some untidy missing memory free() calls in pcretest, to keep
3086 Electric Fence happy when testing.
3087
3088
3089
3090 Version 5.0 13-Sep-04
3091 ---------------------
3092
3093 1. Internal change: literal characters are no longer packed up into items
3094 containing multiple characters in a single byte-string. Each character
3095 is now matched using a separate opcode. However, there may be more than one
3096 byte in the character in UTF-8 mode.
3097
3098 2. The pcre_callout_block structure has two new fields: pattern_position and
3099 next_item_length. These contain the offset in the pattern to the next match
3100 item, and its length, respectively.
3101
3102 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3103 insertion of callouts before each pattern item. Added the /C option to
3104 pcretest to make use of this.
3105
3106 4. On the advice of a Windows user, the lines
3107
3108 #if defined(_WIN32) || defined(WIN32)
3109 _setmode( _fileno( stdout ), 0x8000 );
3110 #endif /* defined(_WIN32) || defined(WIN32) */
3111
3112 have been added to the source of pcretest. This apparently does useful
3113 magic in relation to line terminators.
3114
3115 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3116 for the benefit of those environments where the "b" makes a difference.
3117
3118 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3119 to know about it. I have put a hack into configure.in that adds in code
3120 to set GCC=yes if CC=icc. This seems to end up at a point in the
3121 generated configure script that is early enough to affect the setting of
3122 compiler options, which is what is needed, but I have no means of testing
3123 whether it really works. (The user who reported this had patched the
3124 generated configure script, which of course I cannot do.)
3125
3126 LATER: After change 22 below (new libtool files), the configure script
3127 seems to know about icc (and also ecc). Therefore, I have commented out
3128 this hack in configure.in.
3129
3130 7. Added support for pkg-config (2 patches were sent in).
3131
3132 8. Negated POSIX character classes that used a combination of internal tables
3133 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3134 [[:^ascii]]. Typically, they would match almost any characters. The other
3135 POSIX classes were not broken in this way.
3136
3137 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3138 to find the match, as PCRE was deluded into thinking that the match had to
3139 start at the start point or following a newline. The same bug applied to
3140 patterns with negative forward assertions or any backward assertions
3141 preceding ".*" at the start, unless the pattern required a fixed first
3142 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3143
3144 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3145 starting at the last subject character, bytes beyond the end of the subject
3146 string were read.
3147
3148 11. Renamed the variable "class" as "classbits" to make life easier for C++
3149 users. (Previously there was a macro definition, but it apparently wasn't
3150 enough.)
3151
3152 12. Added the new field "tables" to the extra data so that tables can be passed
3153 in at exec time, or the internal tables can be re-selected. This allows
3154 a compiled regex to be saved and re-used at a later time by a different
3155 program that might have everything at different addresses.
3156
3157 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3158 -R library as well as a -L library.
3159
3160 14. The debugging options of pcretest (-d on the command line or D on a
3161 pattern) showed incorrect output for anything following an extended class
3162 that contained multibyte characters and which was followed by a quantifier.
3163
3164 15. Added optional support for general category Unicode character properties
3165 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3166 support. It adds about 90K to the size of the library. The meanings of the
3167 inbuilt class escapes such as \d and \s have NOT been changed.
3168
3169 16. Updated pcredemo.c to include calls to free() to release the memory for the
3170 compiled pattern.
3171
3172 17. The generated file chartables.c was being created in the source directory
3173 instead of in the building directory. This caused the build to fail if the
3174 source directory was different from the building directory, and was
3175 read-only.
3176
3177 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3178 file. No doubt somebody will tell me if they don't make sense... Also added
3179 Dan Mooney's comments about building on OpenVMS.
3180
3181 19. Added support for partial matching via the PCRE_PARTIAL option for
3182 pcre_exec() and the \P data escape in pcretest.
3183
3184 20. Extended pcretest with 3 new pattern features:
3185
3186 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3187 write the compiled pattern to the file whose name is "rest-of-line".
3188 This is a straight binary dump of the data, with the saved pointer to
3189 the character tables forced to be NULL. The study data, if any, is
3190 written too. After writing, pcretest reads a new pattern.
3191
3192 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3193 compiled pattern from the given file. There must not be any
3194 occurrences of "<" in the file name (pretty unlikely); if there are,
3195 pcretest will instead treat the initial "<" as a pattern delimiter.
3196 After reading in the pattern, pcretest goes on to read data lines as
3197 usual.
3198
3199 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3200 and 16-bit fields in a compiled pattern, to simulate a pattern that
3201 was compiled on a host of opposite endianness.
3202
3203 21. The pcre-exec() function can now cope with patterns that were compiled on
3204 hosts of opposite endianness, with this restriction:
3205
3206 As for any compiled expression that is saved and used later, the tables
3207 pointer field cannot be preserved; the extra_data field in the arguments
3208 to pcre_exec() should be used to pass in a tables address if a value
3209 other than the default internal tables were used at compile time.
3210
3211 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3212 now diagnosed as an error. Previously, most of the time, a negative number
3213 would have been treated as zero, but if in addition "ovector" was passed as
3214 NULL, a crash could occur.
3215
3216 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3217 new versions from the libtool 1.5 distribution (the last one is a copy of
3218 a file called libtool.m4). This seems to have fixed the need to patch
3219 "configure" to support Darwin 1.3 (which I used to do). However, I still
3220 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3221 workstation).
3222
3223 24. Changed the PCRE licence to be the more standard "BSD" licence.
3224
3225
3226 Version 4.5 01-Dec-03
3227 ---------------------
3228
3229 1. There has been some re-arrangement of the code for the match() function so
3230 that it can be compiled in a version that does not call itself recursively.
3231 Instead, it keeps those local variables that need separate instances for
3232 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3233 needs to "recurse". Keeping track of where control must go is done by means
3234 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3235 hide most of the details from the main code, and operates only if
3236 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3237 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3238 operating.
3239
3240 To make it easier for callers to provide specially tailored get/free
3241 functions for this usage, two new functions, pcre_stack_malloc, and
3242 pcre_stack_free, are used. They are always called in strict stacking order,
3243 and the size of block requested is always the same.
3244
3245 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3246 PCRE has been compiled to use the stack or the heap for recursion. The
3247 -C option of pcretest uses this to show which version is compiled.
3248
3249 A new data escape \S, is added to pcretest; it causes the amounts of store
3250 obtained and freed by both kinds of malloc/free at match time to be added
3251 to the output.
3252
3253 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3254 what's available on my current Linux desktop machine.
3255
3256 3. When matching a UTF-8 string, the test for a valid string at the start has
3257 been extended. If start_offset is not zero, PCRE now checks that it points
3258 to a byte that is the start of a UTF-8 character. If not, it returns
3259 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3260 this is necessary because there may be backward assertions in the pattern.
3261 When matching the same subject several times, it may save resources to use
3262 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3263
3264 4. The code for checking the validity of UTF-8 strings has been tightened so
3265 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3266 containing "overlong sequences".
3267
3268 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3269 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3270 should have been "&", but it just so happened that all the cases this let
3271 through by mistake were picked up later in the function.
3272
3273 6. I had used a variable called "isblank" - this is a C99 function, causing
3274 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3275
3276 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3277 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3278 so the version is shown; (c) stop "make test" echoing "./RunTest".
3279
3280 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3281
3282 9. The prototype for memmove() for systems that don't have it was using
3283 size_t, but the inclusion of the header that defines size_t was later. I've
3284 moved the #includes for the C headers earlier to avoid this.
3285
3286 10. Added some adjustments to the code to make it easier to compiler on certain
3287 special systems:
3288
3289 (a) Some "const" qualifiers were missing.
3290 (b) Added the macro EXPORT before all exported functions; by default this
3291 is defined to be empty.
3292 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3293 that it reads its output file name as an argument instead of writing
3294 to the standard output and assuming this can be redirected.
3295
3296 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3297 class containing characters with values greater than 255, PCRE compilation
3298 went into a loop.
3299
3300 12. A recursive reference to a subpattern that was within another subpattern
3301 that had a minimum quantifier of zero caused PCRE to crash. For example,
3302 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3303 recursion. If the recursively-called subpattern itself had a zero repeat,
3304 that was OK.
3305
3306 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3307 buffer into which it was copied (for escape processing) was still set at
3308 1024, so long lines caused crashes.
3309
3310 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3311 "internal error: code overflow...". This applied to any character class
3312 that was followed by a possessive quantifier.
3313
3314 15. Modified the Makefile to add libpcre.la as a prerequisite for
3315 libpcreposix.la because I was told this is needed for a parallel build to
3316 work.
3317
3318 16. If a pattern that contained .* following optional items at the start was
3319 studied, the wrong optimizing data was generated, leading to matching
3320 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3321 matching string must start with a or b or c. The correct conclusion for
3322 this pattern is that a match can start with any character.
3323
3324
3325 Version 4.4 13-Aug-03
3326 ---------------------
3327
3328 1. In UTF-8 mode, a character class containing characters with values between
3329 127 and 255 was not handled correctly if the compiled pattern was studied.
3330 In fixing this, I have also improved the studying algorithm for such
3331 classes (slightly).
3332
3333 2. Three internal functions had redundant arguments passed to them. Removal
3334 might give a very teeny performance improvement.
3335
3336 3. Documentation bug: the value of the capture_top field in a callout is *one
3337 more than* the number of the hightest numbered captured substring.
3338
3339 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3340 in incorrectly linking with a previously installed version. They now link
3341 explicitly with libpcre.la.
3342
3343 5. configure.in no longer needs to recognize Cygwin specially.
3344
3345 6. A problem in pcre.in for Windows platforms is fixed.
3346
3347 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3348 pcretest, it used to include the size of the study block as part of its
3349 output. Unfortunately, the structure contains a field that has a different
3350 size on different hardware architectures. This meant that the tests that
3351 showed this size failed. As the block is currently always of a fixed size,
3352 this information isn't actually particularly useful in pcretest output, so
3353 I have just removed it.
3354
3355 8. Three pre-processor statements accidentally did not start in column 1.
3356 Sadly, there are *still* compilers around that complain, even though
3357 standard C has not required this for well over a decade. Sigh.
3358
3359 9. In pcretest, the code for checking callouts passed small integers in the
3360 callout_data field, which is a void * field. However, some picky compilers
3361 complained about the casts involved for this on 64-bit systems. Now
3362 pcretest passes the address of the small integer instead, which should get
3363 rid of the warnings.
3364
3365 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3366 both compile and run time, and gives an error if an invalid UTF-8 sequence
3367 is found. There is a option for disabling this check in cases where the
3368 string is known to be correct and/or the maximum performance is wanted.
3369
3370 11. In response to a bug report, I changed one line in Makefile.in from
3371
3372 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3373 to
3374 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3375
3376 to look similar to other lines, but I have no way of telling whether this
3377 is the right thing to do, as I do not use Windows. No doubt I'll get told
3378 if it's wrong...
3379
3380
3381 Version 4.3 21-May-03
3382 ---------------------
3383
3384 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3385 Makefile.
3386
3387 2. Some refactoring to improve the quality of the code:
3388
3389 (i) The utf8_table... variables are now declared "const".
3390
3391 (ii) The code for \cx, which used the "case flipping" table to upper case
3392 lower case letters, now just substracts 32. This is ASCII-specific,
3393 but the whole concept of \cx is ASCII-specific, so it seems
3394 reasonable.
3395
3396 (iii) PCRE was using its character types table to recognize decimal and
3397 hexadecimal digits in the pattern. This is silly, because it handles
3398 only 0-9, a-f, and A-F, but the character types table is locale-
3399 specific, which means strange things might happen. A private
3400 table is now used for this - though it costs 256 bytes, a table is
3401 much faster than multiple explicit tests. Of course, the standard
3402 character types table is still used for matching digits in subject
3403 strings against \d.
3404
3405 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3406 ending in _t are). So I've renamed it as ESC_tee.
3407
3408 3. The first argument for regexec() in the POSIX wrapper should have been
3409 defined as "const".
3410
3411 4. Changed pcretest to use malloc() for its buffers so that they can be
3412 Electric Fenced for debugging.
3413
3414 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3415 to read one or more bytes before the start of the subject string. Often this
3416 had no effect on PCRE's behaviour, but in some circumstances it could
3417 provoke a segmentation fault.
3418
3419 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3420 to try to read one or more bytes before the start of the subject string.
3421
3422 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3423 UTF-8 support could misbehave in various ways if the subject string
3424 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3425 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3426 back over UTF-8 characters.)
3427
3428
3429 Version 4.2 14-Apr-03
3430 ---------------------
3431
3432 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3433
3434 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3435 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3436 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3437 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3438 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3439 and BUILD_EXEEXT
3440 Note: automatic setting of the BUILD variables is not yet working
3441 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3442 compile-time but not at link-time
3443 [LINK]: use for linking executables only
3444 make different versions for Windows and non-Windows
3445 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3446 libraries
3447 [LINK_FOR_BUILD]: new variable
3448 [OBJEXT]: use throughout
3449 [EXEEXT]: use throughout
3450 <winshared>: new target
3451 <wininstall>: new target
3452 <dftables.o>: use native compiler
3453 <dftables>: use native linker
3454 <install>: handle Windows platform correctly
3455 <clean>: ditto
3456 <check>: ditto
3457 copy DLL to top builddir before testing
3458
3459 As part of these changes, -no-undefined was removed again. This was reported
3460 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3461 in any case.
3462
3463 3. Some tidies to get rid of compiler warnings:
3464
3465 . In the match_data structure, match_limit was an unsigned long int, whereas
3466 match_call_count was an int. I've made them both unsigned long ints.
3467
3468 . In pcretest the fact that a const uschar * doesn't automatically cast to
3469 a void * provoked a warning.
3470
3471 . Turning on some more compiler warnings threw up some "shadow" variables
3472 and a few more missing casts.
3473
3474 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3475 option, a class that contained a single character with a value between 128
3476 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3477
3478 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3479 option, a class that contained several characters, but with at least one
3480 whose value was between 128 and 255 caused PCRE to crash.
3481
3482
3483 Version 4.1 12-Mar-03
3484 ---------------------
3485
3486 1. Compiling with gcc -pedantic found a couple of places where casts were
3487 needed, and a string in dftables.c that was longer than standard compilers are
3488 required to support.
3489
3490 2. Compiling with Sun's compiler found a few more places where the code could
3491 be tidied up in order to avoid warnings.
3492
3493 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3494 first of these names is deprecated in the latest Autoconf in favour of the name
3495 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3496 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3497 analogy I have changed it to CFLAGS_FOR_BUILD.
3498
3499 4. Added -no-undefined to the linking command in the Makefile, because this is
3500 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3501 linking step for the pcreposix library.
3502
3503 5. PCRE was failing to diagnose the case of two named groups with the same
3504 name.
3505
3506 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3507 literal character that is needed in the subject for a match, and scans along to
3508 ensure that it is present before embarking on the full matching process. This
3509 saves time in cases of nested unlimited repeats that are never going to match.
3510 Problem: the scan can take a lot of time if the subject is very long (e.g.
3511 megabytes), thus penalizing straightforward matches. It is now done only if the
3512 amount of subject to be scanned is less than 1000 bytes.
3513
3514 7. A lesser problem with the same optimization is that it was recording the
3515 first character of an anchored pattern as "needed", thus provoking a search
3516 right along the subject, even when the first match of the pattern was going to
3517 fail. The "needed" character is now not set for anchored patterns, unless it
3518 follows something in the pattern that is of non-fixed length. Thus, it still
3519 fulfils its original purpose of finding quick non-matches in cases of nested
3520 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3521
3522
3523 Version 4.0 17-Feb-03
3524 ---------------------
3525
3526 1. If a comment in an extended regex that started immediately after a meta-item
3527 extended to the end of string, PCRE compiled incorrect data. This could lead to
3528 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3529
3530 2. Moved to autoconf 2.53 and libtool 1.4.2.
3531
3532 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3533 the special perltest8 script is no longer needed - all the tests can be run
3534 from a single perltest script.
3535
3536 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3537 by \s. It has now been removed in PCRE. This means it isn't recognized as
3538 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3539 class [:space:] *does* include VT, thereby creating a mess.
3540
3541 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3542 space and tab.
3543
3544 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3545 its new features into the main test script, reducing the number of scripts.
3546
3547 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3548 were backward compatible, and made the (?i) apply to the whole pattern, as if
3549 /i were given. Now it behaves more logically, and applies the option setting
3550 only to what follows. PCRE has been changed to follow suit. However, if it
3551 finds options settings right at the start of the pattern, it extracts them into
3552 the global options, as before. Thus, they show up in the info data.
3553
3554 8. Added support for the \Q...\E escape sequence. Characters in between are
3555 treated as literals. This is slightly different from Perl in that $ and @ are
3556 also handled as literals inside the quotes. In Perl, they will cause variable
3557 interpolation. Note the following examples:
3558
3559 Pattern PCRE matches Perl matches
3560
3561 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3562 \Qabc\$xyz\E abc\$xyz abc\$xyz
3563 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3564
3565 For compatibility with Perl, \Q...\E sequences are recognized inside character
3566 classes as well as outside them.
3567
3568 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3569 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3570 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3571 signed/unsigned warnings.
3572
3573 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3574 option for pcretest, so I've replaced it by a simple function that does just
3575 that job.
3576
3577 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3578 "pcregrep -".
3579
3580 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3581 Java package. This provides some syntactic sugar for simple cases of what my
3582 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3583 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3584 item, you can use this simplified notation. Note that only makes sense with
3585 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3586 greediness, whatever the setting of the PCRE_UNGREEDY option.
3587
3588 13. A change of greediness default within a pattern was not taking effect at
3589 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3590 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3591 was abstracted outside.
3592
3593 14. PCRE now supports the \G assertion. It is true when the current matching
3594 position is at the start point of the match. This differs from \A when the
3595 starting offset is non-zero. Used with the /g option of pcretest (or similar
3596 code), it works in the same way as it does for Perl's /g option. If all
3597 alternatives of a regex begin with \G, the expression is anchored to the start
3598 match position, and the "anchored" flag is set in the compiled expression.
3599
3600 15. Some bugs concerning the handling of certain option changes within patterns
3601 have been fixed. These applied to options other than (?ims). For example,
3602 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3603 been the other way round. Some of this was related to change 7 above.
3604
3605 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3606 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3607 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3608 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3609
3610 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3611 mode. Unlike ".", it always matches newline, whatever the setting of
3612 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3613 assertions. Perl allows it, but it doesn't (in general) work because it can't
3614 calculate the length of the lookbehind. At least, that's the case for Perl
3615 5.8.0 - I've been told they are going to document that it doesn't work in
3616 future.
3617
3618 18. Added an error diagnosis for escapes that PCRE does not support: these are
3619 \L, \l, \N, \P, \p, \U, \u, and \X.
3620
3621 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3622 reading past the end of the pattern in cases such as /[abcd/.
3623
3624 20. PCRE was getting more memory than necessary for patterns with classes that
3625 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3626
3627 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3628 compiling PCRE for use with Virtual Pascal.
3629
3630 22. Small fix to the Makefile to make it work properly if the build is done
3631 outside the source tree.
3632
3633 23. Added a new extension: a condition to go with recursion. If a conditional
3634 subpattern starts with (?(R) the "true" branch is used if recursion has
3635 happened, whereas the "false" branch is used only at the top level.
3636
3637 24. When there was a very long string of literal characters (over 255 bytes
3638 without UTF support, over 250 bytes with UTF support), the computation of how
3639 much memory was required could be incorrect, leading to segfaults or other
3640 strange effects.
3641
3642 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3643 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3644 there was a subsequent back reference to those brackets. This meant that, for
3645 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3646 possible to check for precisely this case. All we can do is abandon the
3647 optimization if .* occurs inside capturing brackets when there are any back
3648 references whatsoever. (See below for a better fix that came later.)
3649
3650 26. The handling of the optimization for finding the first character of a
3651 non-anchored pattern, and for finding a character that is required later in the
3652 match were failing in some cases. This didn't break the matching; it just
3653 failed to optimize when it could. The way this is done has been re-implemented.
3654
3655 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3656
3657 28. Added a new feature that provides some of the functionality that Perl
3658 provides with (?{...}). The facility is termed a "callout". The way it is done
3659 in PCRE is for the caller to provide an optional function, by setting
3660 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3661 global variable. By default it is unset, which disables all calling out. To get
3662 the function called, the regex must include (?C) at appropriate points. This
3663 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3664 This provides a means of identifying different callout points. When PCRE
3665 reaches such a point in the regex, if pcre_callout has been set, the external
3666 function is called. It is provided with data in a structure called
3667 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3668 matching continues; if it returns a non-zero value, the match at the current
3669 point fails. However, backtracking will occur if possible. [This was changed
3670 later and other features added - see item 49 below.]
3671
3672 29. pcretest is upgraded to test the callout functionality. It provides a
3673 callout function that displays information. By default, it shows the start of
3674 the match and the current position in the text. There are some new data escapes
3675 to vary what happens:
3676
3677 \C+ in addition, show current contents of captured substrings
3678 \C- do not supply a callout function
3679 \C!n return 1 when callout number n is reached
3680 \C!n!m return 1 when callout number n is reached for the mth time
3681
3682 30. If pcregrep was called with the -l option and just a single file name, it
3683 output "<stdin>" if a match was found, instead of the file name.
3684
3685 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3686 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3687 pcre_exec(). This saves a malloc/free per call. The default value of
3688 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3689 when configuring.
3690
3691 32. The default maximum size of a compiled pattern is 64K. There have been a
3692 few cases of people hitting this limit. The code now uses macros to handle the
3693 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3694 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3695 configuring. Tests 2 and 5 work only with 2-byte links because they output
3696 debugging information about compiled patterns.
3697
3698 33. Internal code re-arrangements:
3699
3700 (a) Moved the debugging function for printing out a compiled regex into
3701 its own source file (printint.c) and used #include to pull it into
3702 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3703 separate copies.
3704
3705 (b) Defined the list of op-code names for debugging as a macro in
3706 internal.h so that it is next to the definition of the opcodes.
3707
3708 (c) Defined a table of op-code lengths for simpler skipping along compiled
3709 code. This is again a macro in internal.h so that it is next to the
3710 definition of the opcodes.
3711
3712 34. Added support for recursive calls to individual subpatterns, along the
3713 lines of Robin Houston's patch (but implemented somewhat differently).
3714
3715 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3716 allow it to read and process whole directories in Win32. This code was
3717 contributed by Lionel Fourquaux; it has not been tested by me.
3718
3719 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3720 used to name a group. Names consist of alphanumerics and underscores, and must
3721 be unique. Back references use the syntax (?P=name) and recursive calls use
3722 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3723 numbers. The function pcre_fullinfo() can be used after compilation to extract
3724 a name/number map. There are three relevant calls:
3725
3726 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3727 PCRE_INFO_NAMECOUNT yields the number of entries
3728 PCRE_INFO_NAMETABLE yields a pointer to the map.
3729
3730 The map is a vector of fixed-size entries. The size of each entry depends on
3731 the length of the longest name used. The first two bytes of each entry are the
3732 group number, most significant byte first. There follows the corresponding
3733 name, zero terminated. The names are in alphabetical order.
3734
3735 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3736 case instead of 255. Making it the same both with and without UTF-8 support
3737 means that the same test output works with both.
3738
3739 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3740 calling malloc() with a zero argument.
3741
3742 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3743 optimization. I've improved things by keeping a bitmap of backreferences with
3744 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3745 fact referenced, the optimization can be applied. It is unlikely that a
3746 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3747 the match to follow \n) will appear inside brackets with a number greater than
3748 31, but if it does, any back reference > 31 suppresses the optimization.
3749
3750 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3751 of disabling numbered capturing parentheses. Any opening parenthesis that is
3752 not followed by ? behaves as if it were followed by ?: but named parentheses
3753 can still be used for capturing (and they will acquire numbers in the usual
3754 way).
3755
3756 41. Redesigned the return codes from the match() function into yes/no/error so
3757 that errors can be passed back from deep inside the nested calls. A malloc
3758 failure while inside a recursive subpattern call now causes the
3759 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3760
3761 42. It is now possible to set a limit on the number of times the match()
3762 function is called in a call to pcre_exec(). This facility makes it possible to
3763 limit the amount of recursion and backtracking, though not in a directly
3764 obvious way, because the match() function is used in a number of different
3765 circumstances. The count starts from zero for each position in the subject
3766 string (for non-anchored patterns). The default limit is, for compatibility, a
3767 large number, namely 10 000 000. You can change this in two ways:
3768
3769 (a) When configuring PCRE before making, you can use --with-match-limit=n
3770 to set a default value for the compiled library.
3771
3772 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3773 a different value is set. See 45 below.
3774
3775 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3776
3777 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3778 of things that can be changed at compile time. The first argument specifies
3779 what is wanted and the second points to where the information is to be placed.
3780 The current list of available information is:
3781
3782 PCRE_CONFIG_UTF8
3783
3784 The output is an integer that is set to one if UTF-8 support is available;
3785 otherwise it is set to zero.
3786
3787 PCRE_CONFIG_NEWLINE
3788
3789 The output is an integer that it set to the value of the code that is used for
3790 newline. It is either LF (10) or CR (13).
3791
3792 PCRE_CONFIG_LINK_SIZE
3793
3794 The output is an integer that contains the number of bytes used for internal
3795 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3796
3797 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3798
3799 The output is an integer that contains the threshold above which the POSIX
3800 interface uses malloc() for output vectors. See item 31 above.
3801
3802 PCRE_CONFIG_MATCH_LIMIT
3803
3804 The output is an unsigned integer that contains the default limit of the number
3805 of match() calls in a pcre_exec() execution. See 42 above.
3806
3807 44. pcretest has been upgraded by the addition of the -C option. This causes it
3808 to extract all the available output from the new pcre_config() function, and to
3809 output it. The program then exits immediately.
3810
3811 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3812 order to support additional features. One way would have been to define
3813 pcre_exec2() (for example) with extra arguments, but this would not have been
3814 extensible, and would also have required all calls to the original function to
3815 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3816 is used for passing in "extra" data from pcre_study().
3817
3818 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3819 contains the following fields:
3820
3821 flags a bitmap indicating which of the following fields are set
3822 study_data opaque data from pcre_study()
3823 match_limit a way of specifying a limit on match() calls for a specific
3824 call to pcre_exec()
3825 callout_data data for callouts (see 49 below)
3826
3827 The flag bits are also defined in pcre.h, and are
3828
3829 PCRE_EXTRA_STUDY_DATA
3830 PCRE_EXTRA_MATCH_LIMIT
3831 PCRE_EXTRA_CALLOUT_DATA
3832
3833 The pcre_study() function now returns one of these new pcre_extra blocks, with
3834 the actual study data pointed to by the study_data field, and the
3835 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3836 before. That is, this change is entirely upwards-compatible and requires no
3837 change to existing code.
3838
3839 If you want to pass in additional data to pcre_exec(), you can either place it
3840 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3841 block.
3842
3843 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3844 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3845 times with different match limits, until it finds the minimum value needed for
3846 pcre_exec() to complete. The value is then output. This can be instructive; for
3847 most simple matches the number is quite small, but for pathological cases it
3848 gets very large very quickly.
3849
3850 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3851 returns the size of the data block pointed to by the study_data field in a
3852 pcre_extra block, that is, the value that was passed as the argument to
3853 pcre_malloc() when PCRE was getting memory in which to place the information
3854 created by pcre_study(). The fourth argument should point to a size_t variable.
3855 pcretest has been extended so that this information is shown after a successful
3856 pcre_study() call when information about the compiled regex is being displayed.
3857
3858 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3859 because what follows is always an absolute path. (Later: it turns out that this
3860 is more than cosmetic for MinGW, because it doesn't like empty path
3861 components.)
3862
3863 49. Some changes have been made to the callout feature (see 28 above):
3864
3865 (i) A callout function now has three choices for what it returns:
3866
3867 0 => success, carry on matching
3868 > 0 => failure at this point, but backtrack if possible
3869 < 0 => serious error, return this value from pcre_exec()
3870
3871 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3872 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3873 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3874 use by callout functions. It will never be used by PCRE itself.
3875
3876 (ii) The pcre_extra structure (see 45 above) has a void * field called
3877 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3878 pcre_callout_block structure has a field of the same name. The contents of
3879 the field passed in the pcre_extra structure are passed to the callout
3880 function in the corresponding field in the callout block. This makes it
3881 easier to use the same callout-containing regex from multiple threads. For
3882 testing, the pcretest program has a new data escape
3883
3884 \C*n pass the number n (may be negative) as callout_data
3885
3886 If the callout function in pcretest receives a non-zero value as
3887 callout_data, it returns that value.
3888
3889 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3890 there were some redundant $(CFLAGS) in commands that are now specified as
3891 $(LINK), which already includes $(CFLAGS).
3892
3893 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3894 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3895 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3896 one-byte characters throughout. Note that case-insensitive matching applies
3897 only to characters whose values are less than 256. PCRE doesn't support the
3898 notion of cases for higher-valued characters.
3899
3900 (i) A character class whose characters are all within 0-255 is handled as
3901 a bit map, and the map is inverted for negative classes. Previously, a
3902 character > 255 always failed to match such a class; however it should
3903 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3904
3905 (ii) A negated character class with a single character < 255 is coded as
3906 "not this character" (OP_NOT). This wasn't working properly when the test
3907 character was multibyte, either singly or repeated.
3908
3909 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3910 mode, for example: \x{100}{2,3}.
3911
3912 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3913 singly or repeated) now correctly test multibyte characters. However,
3914 PCRE doesn't recognize any characters with values greater than 255 as
3915 digits, spaces, or word characters. Such characters always match \D, \S,
3916 and \W, and never match \d, \s, or \w.
3917
3918 (v) Classes may now contain characters and character ranges with values
3919 greater than 255. For example: [ab\x{100}-\x{400}].
3920
3921 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3922 PCRE in UTF-8 mode.
3923
3924 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3925 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3926 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3927 value.)
3928
3929 53. The single man page has become too large. I have therefore split it up into
3930 a number of separate man pages. These also give rise to individual HTML pages;
3931 these are now put in a separate directory, and there is an index.html page that
3932 lists them all. Some hyperlinking between the pages has been installed.
3933
3934 54. Added convenience functions for handling named capturing parentheses.
3935
3936 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3937 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3938 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3939 are faulted.
3940
3941 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3942 calling configure. These values are used when compiling the dftables.c program
3943 which is run to generate the source of the default character tables. They
3944 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3945 you will need to set these values.
3946
3947 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3948
3949
3950 Version 3.9 02-Jan-02
3951 ---------------------
3952
3953 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3954
3955 2. If --disable-static was given, the building process failed when trying to
3956 build pcretest and pcregrep. (For some reason it was using libtool to compile
3957 them, which is not right, as they aren't part of the library.)
3958
3959
3960 Version 3.8 18-Dec-01
3961 ---------------------
3962
3963 1. The experimental UTF-8 code was completely screwed up. It was packing the
3964 bytes in the wrong order. How dumb can you get?
3965
3966
3967 Version 3.7 29-Oct-01
3968 ---------------------
3969
3970 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3971 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3972 this didn't happen under Solaris 8, where I normally test things.
3973
3974 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3975 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3976 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3977 this fix an hour or so after the initial 3.7 release.)
3978
3979
3980 Version 3.6 23-Oct-01
3981 ---------------------
3982
3983 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3984 offsets passed as NULL with zero offset count.
3985
3986 2. The config.guess and config.sub files had not been updated when I moved to
3987 the latest autoconf.
3988
3989
3990 Version 3.5 15-Aug-01
3991 ---------------------
3992
3993 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3994 had been forgotten.
3995
3996 2. By using declared but undefined structures, we can avoid using "void"
3997 definitions in pcre.h while keeping the internal definitions of the structures
3998 private.
3999
4000 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4001 user point of view, this means that both static and shared libraries are built
4002 by default, but this can be individually controlled. More of the work of
4003 handling this static/shared cases is now inside libtool instead of PCRE's make
4004 file.
4005
4006 4. The pcretest utility is now installed along with pcregrep because it is
4007 useful for users (to test regexs) and by doing this, it automatically gets
4008 relinked by libtool. The documentation has been turned into a man page, so
4009 there are now .1, .txt, and .html versions in /doc.
4010
4011 5. Upgrades to pcregrep:
4012 (i) Added long-form option names like gnu grep.
4013 (ii) Added --help to list all options with an explanatory phrase.
4014 (iii) Added -r, --recursive to recurse into sub-directories.
4015 (iv) Added -f, --file to read patterns from a file.
4016
4017 6. pcre_exec() was referring to its "code" argument before testing that
4018 argument for NULL (and giving an error if it was NULL).
4019
4020 7. Upgraded Makefile.in to allow for compiling in a different directory from
4021 the source directory.
4022
4023 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4024 options bits, the pointer it was passed was to an int instead of to an unsigned
4025 long int. This mattered only on 64-bit systems.
4026
4027 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4028 generated) instead of pcre.in, which it its source. Also made the same change
4029 in several of the .c files.
4030
4031 10. A new release of gcc defines printf() as a macro, which broke pcretest
4032 because it had an ifdef in the middle of a string argument for printf(). Fixed
4033 by using separate calls to printf().
4034
4035 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4036 script, to force use of CR or LF instead of \n in the source. On non-Unix
4037 systems, the value can be set in config.h.
4038
4039 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4040 absolute limit. Changed the text of the error message to make this clear, and
4041 likewise updated the man page.
4042
4043 13. The limit of 99 on the number of capturing subpatterns has been removed.
4044 The new limit is 65535, which I hope will not be a "real" limit.
4045
4046
4047 Version 3.4 22-Aug-00
4048 ---------------------
4049
4050 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4051
4052 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4053
4054
4055 Version 3.3 01-Aug-00
4056 ---------------------
4057
4058 1. If an octal character was given, but the value was greater than \377, it
4059 was not getting masked to the least significant bits, as documented. This could
4060 lead to crashes in some systems.
4061
4062 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4063 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4064
4065 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4066 These just pass their arguments on to (pcre_free)(), but they are provided
4067 because some uses of PCRE bind it to non-C systems that can call its functions,
4068 but cannot call free() or pcre_free() directly.
4069
4070 4. Add "make test" as a synonym for "make check". Corrected some comments in
4071 the Makefile.
4072
4073 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4074 Makefile.
4075
4076 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4077 command called pgrep for grepping around the active processes.
4078
4079 7. Added the beginnings of support for UTF-8 character strings.
4080
4081 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4082 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4083 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4084 out for the ar command.)
4085
4086
4087 Version 3.2 12-May-00
4088 ---------------------
4089
4090 This is purely a bug fixing release.
4091
4092 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4093 of ZA. This was just one example of several cases that could provoke this bug,
4094 which was introduced by change 9 of version 2.00. The code for breaking
4095 infinite loops after an iteration that matches an empty string was't working
4096 correctly.
4097
4098 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4099 when matched against abbab (for example). After matching an empty string, it
4100 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4101 caused it to match further down the string than it should.
4102
4103 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4104 was there because it doesn't seem to be needed, and it causes trouble on some
4105 systems, as it is not a Standard C header. It has been removed.
4106
4107 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4108 were reported on the Macintosh. The changes were from
4109
4110 while ((c = *(++ptr)) != 0 && c != '\n');
4111 to
4112 while ((c = *(++ptr)) != 0 && c != '\n') ;
4113
4114 Totally extraordinary, but if that's what it takes...
4115
4116 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4117 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4118 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4119 assumes the way PCRE uses memmove() (always moving upwards).
4120
4121 6. PCRE is being used in one environment where strchr() is not available. There
4122 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4123 faster code anyway.
4124
4125
4126 Version 3.1 09-Feb-00
4127 ---------------------
4128
4129 The only change in this release is the fixing of some bugs in Makefile.in for
4130 the "install" target:
4131
4132 (1) It was failing to install pcreposix.h.
4133
4134 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4135
4136
4137 Version 3.0 01-Feb-00
4138 ---------------------
4139
4140 1. Add support for the /+ modifier to perltest (to output $` like it does in
4141 pcretest).
4142
4143 2. Add support for the /g modifier to perltest.
4144
4145 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4146 matches null strings.
4147
4148 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4149 pattern. Perl treats empty patterns specially - it reuses the most recent
4150 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4151 effect.
4152
4153 5. The POSIX interface was broken in that it was just handing over the POSIX
4154 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4155 required a bigger vector, with some working space on the end. This means that
4156 the POSIX wrapper now has to get and free some memory, and copy the results.
4157
4158 6. Added some simple autoconf support, placing the test data and the
4159 documentation in separate directories, re-organizing some of the
4160 information files, and making it build pcre-config (a GNU standard). Also added
4161 libtool support for building PCRE as a shared library, which is now the
4162 default.
4163
4164 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4165 09 are not valid octal constants. Single digits will be used for minor values
4166 less than 10.
4167
4168 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4169 existing programs that set these in the POSIX interface can use PCRE without
4170 modification.
4171
4172 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4173 return all that pcre_info() returns, plus additional data. The pcre_info()
4174 function is retained for compatibility, but is considered to be obsolete.
4175
4176 10. Added experimental recursion feature (?R) to handle one common case that
4177 Perl 5.6 will be able to do with (?p{...}).
4178
4179 11. Added support for POSIX character classes like [:alpha:], which Perl is
4180 adopting.
4181
4182
4183 Version 2.08 31-Aug-99
4184 ----------------------
4185
4186 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4187 trying to match at the startoffset position, but instead was moving forward to
4188 the next newline as if a previous match had failed.
4189
4190 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4191 and could get into a loop if a null string was matched other than at the start
4192 of the subject.
4193
4194 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4195 be distinguished at compile time, and for completeness also added PCRE_DATE.
4196
4197 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4198 in GnuWin32 environments.
4199
4200
4201 Version 2.07 29-Jul-99
4202 ----------------------
4203
4204 1. The documentation is now supplied in plain text form and HTML as well as in
4205 the form of man page sources.
4206
4207 2. C++ compilers don't like assigning (void *) values to other pointer types.
4208 In particular this affects malloc(). Although there is no problem in Standard
4209 C, I've put in casts to keep C++ compilers happy.
4210
4211 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4212 should be (const char *).
4213
4214 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4215 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4216 However, I haven't made this a standard facility. The documentation doesn't
4217 mention it, and the Makefile doesn't support it.
4218
4219 5. The Makefile now contains an "install" target, with editable destinations at
4220 the top of the file. The pcretest program is not installed.
4221
4222 6. pgrep -V now gives the PCRE version number and date.
4223
4224 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4225 causing the entire string to be ignored, instead of just the last character.
4226
4227 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4228 non-matching string, it can take a very, very long time, even for strings of
4229 quite modest length, because of the nested recursion. PCRE now does better in
4230 some of these cases. It does this by remembering the last required literal
4231 character in the pattern, and pre-searching the subject to ensure it is present
4232 before running the real match. In other words, it applies a heuristic to detect
4233 some types of certain failure quickly, and in the above example, if presented
4234 with a string that has no trailing " it gives "no match" very quickly.
4235
4236 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4237 other alternatives are tried instead.
4238
4239
4240 Version 2.06 09-Jun-99
4241 ----------------------
4242
4243 1. Change pcretest's output for amount of store used to show just the code
4244 space, because the remainder (the data block) varies in size between 32-bit and
4245 64-bit systems.
4246
4247 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4248 start matching at. This allows lookbehinds to work when searching for multiple
4249 occurrences in a string.
4250
4251 3. Added additional options to pcretest for testing multiple occurrences:
4252
4253 /+ outputs the rest of the string that follows a match
4254 /g loops for multiple occurrences, using the new startoffset argument
4255 /G loops for multiple occurrences by passing an incremented pointer
4256
4257 4. PCRE wasn't doing the "first character" optimization for patterns starting
4258 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4259 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4260 the letter 'x'. On long subject strings, this gives a significant speed-up.
4261
4262
4263 Version 2.05 21-Apr-99
4264 ----------------------
4265
4266 1. Changed the type of magic_number from int to long int so that it works
4267 properly on 16-bit systems.
4268
4269 2. Fixed a bug which caused patterns starting with .* not to work correctly
4270 when the subject string contained newline characters. PCRE was assuming
4271 anchoring for such patterns in all cases, which is not correct because .* will
4272 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4273 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4274 must be retried after every newline in the subject.
4275
4276
4277 Version 2.04 18-Feb-99
4278 ----------------------
4279
4280 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4281 computation of the store needed to hold the pattern was incorrect (too large).
4282 If such patterns were nested a few deep, this could multiply and become a real
4283 problem.
4284
4285 2. Added /M option to pcretest to show the memory requirement of a specific
4286 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4287
4288 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4289 compiled in such a way that the backtracking after subsequent failure was
4290 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4291 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4292
4293
4294 Version 2.03 02-Feb-99
4295 ----------------------
4296
4297 1. Fixed typo and small mistake in man page.
4298
4299 2. Added 4th condition (GPL supersedes if conflict) and created separate
4300 LICENCE file containing the conditions.
4301
4302 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4303 Perl, that is the internal \ allows the delimiter to be included in the
4304 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4305 the final delimiter, add \ to the end of the pattern (to test the error).
4306
4307 4. Added the convenience functions for extracting substrings after a successful
4308 match. Updated pcretest to make it able to test these functions.
4309
4310
4311 Version 2.02 14-Jan-99
4312 ----------------------
4313
4314 1. Initialized the working variables associated with each extraction so that
4315 their saving and restoring doesn't refer to uninitialized store.
4316
4317 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4318 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4319 fix the problem.
4320
4321 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4322 calls, and wasn't printing the correct value for compiling calls. Increased the
4323 default value of LOOPREPEAT, and the number of significant figures in the
4324 times.
4325
4326 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4327
4328 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4329 a building problem on Windows NT with a FAT file system.
4330
4331
4332 Version 2.01 21-Oct-98
4333 ----------------------
4334
4335 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4336 to character tables built by pcre_maketables() in the current locale. If NULL
4337 is passed, the default tables are used.
4338
4339
4340 Version 2.00 24-Sep-98
4341 ----------------------
4342
4343 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4344 it any more.
4345
4346 2. Allow quantification of (?>) groups, and make it work correctly.
4347
4348 3. The first character computation wasn't working for (?>) groups.
4349
4350 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4351 end of the subject) and add 5.005's \z, which really does match only at the
4352 very end of the subject.
4353
4354 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4355
4356 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4357 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4358 localized options. All options to pcre_study() were also removed.
4359
4360 7. Add other new features from 5.005:
4361
4362 $(?<= positive lookbehind
4363 $(?<! negative lookbehind
4364 (?imsx-imsx) added the unsetting capability
4365 such a setting is global if at outer level; local otherwise
4366 (?imsx-imsx:) non-capturing groups with option setting
4367 (?(cond)re|re) conditional pattern matching
4368
4369 A backreference to itself in a repeated group matches the previous
4370 captured string.
4371
4372 8. General tidying up of studying (both automatic and via "study")
4373 consequential on the addition of new assertions.
4374
4375 9. As in 5.005, unlimited repeated groups that could match an empty substring
4376 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4377 runtime if any iteration does actually match an empty substring.
4378
4379 10. Include the RunTest script in the distribution.
4380
4381 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4382 discrepancies, some of which were old and were also with respect to 5.004. They
4383 have now been fixed.
4384
4385
4386 Version 1.09 28-Apr-98
4387 ----------------------
4388
4389 1. A negated single character class followed by a quantifier with a minimum
4390 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4391 program crashes, or just wrong answers. This did not apply to negated classes
4392 containing more than one character, or to minima other than one.
4393
4394
4395 Version 1.08 27-Mar-98
4396 ----------------------
4397
4398 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4399
4400 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4401 latter must appear before anything that relies on it in the pattern.
4402
4403
4404 Version 1.07 16-Feb-98
4405 ----------------------
4406
4407 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4408 repeat of a potentially empty string).
4409
4410
4411 Version 1.06 23-Jan-98
4412 ----------------------
4413
4414 1. Added Markus Oberhumer's little patches for C++.
4415
4416 2. Literal strings longer than 255 characters were broken.
4417
4418
4419 Version 1.05 23-Dec-97
4420 ----------------------
4421
4422 1. Negated character classes containing more than one character were failing if
4423 PCRE_CASELESS was set at run time.
4424
4425
4426 Version 1.04 19-Dec-97
4427 ----------------------
4428
4429 1. Corrected the man page, where some "const" qualifiers had been omitted.
4430
4431 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4432 input syntax.
4433
4434 3. Fixed memory leak which occurred when a regex with back references was
4435 matched with an offsets vector that wasn't big enough. The temporary memory
4436 that is used in this case wasn't being freed if the match failed.
4437
4438 4. Tidied pcretest to ensure it frees memory that it gets.
4439
4440 5. Temporary memory was being obtained in the case where the passed offsets
4441 vector was exactly big enough.
4442
4443 6. Corrected definition of offsetof() from change 5 below.
4444
4445 7. I had screwed up change 6 below and broken the rules for the use of
4446 setjmp(). Now fixed.
4447
4448
4449 Version 1.03 18-Dec-97
4450 ----------------------
4451
4452 1. A erroneous regex with a missing opening parenthesis was correctly
4453 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4454 on some systems.
4455
4456 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4457 it was reported that one broken compiler failed on the former because "code" is
4458 also an independent variable.
4459
4460 3. The erroneous regex a[]b caused an array overrun reference.
4461
4462 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4463 fail on data ending with that character. (It was going on too far, and checking
4464 the next character, typically a binary zero.) This was specific to the
4465 optimized code for single-character negative classes.
4466
4467 5. Added a contributed patch from the TIN world which does the following:
4468
4469 + Add an undef for memmove, in case the the system defines a macro for it.
4470
4471 + Add a definition of offsetof(), in case there isn't one. (I don't know
4472 the reason behind this - offsetof() is part of the ANSI standard - but
4473 it does no harm).
4474
4475 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4476 most of the places where whitespace preceded '#'. I have given up and
4477 allowed the remaining 2 cases to be at the margin.
4478
4479 + Rename some variables in pcre to eliminate shadowing. This seems very
4480 pedantic, but does no harm, of course.
4481
4482 6. Moved the call to setjmp() into its own function, to get rid of warnings
4483 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4484
4485 7. Constructs such as \d{8,} were compiling into the equivalent of
4486 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4487 outcome, but in this particular case used more store than had been allocated,
4488 which caused the bug to be discovered because it threw up an internal error.
4489
4490 8. The debugging code in both pcre and pcretest for outputting the compiled
4491 form of a regex was going wrong in the case of back references followed by
4492 curly-bracketed repeats.
4493
4494
4495 Version 1.02 12-Dec-97
4496 ----------------------
4497
4498 1. Typos in pcre.3 and comments in the source fixed.
4499
4500 2. Applied a contributed patch to get rid of places where it used to remove
4501 'const' from variables, and fixed some signed/unsigned and uninitialized
4502 variable warnings.
4503
4504 3. Added the "runtest" target to Makefile.
4505
4506 4. Set default compiler flag to -O2 rather than just -O.
4507
4508
4509 Version 1.01 19-Nov-97
4510 ----------------------
4511
4512 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4513 like /([ab]*)*/, that is, for classes with more than one character in them.
4514
4515 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4516 as /((?>a*))*/ (a PCRE_EXTRA facility).
4517
4518
4519 Version 1.00 18-Nov-97
4520 ----------------------
4521
4522 1. Added compile-time macros to support systems such as SunOS4 which don't have
4523 memmove() or strerror() but have other things that can be used instead.
4524
4525 2. Arranged that "make clean" removes the executables.
4526
4527
4528 Version 0.99 27-Oct-97
4529 ----------------------
4530
4531 1. Fixed bug in code for optimizing classes with only one character. It was
4532 initializing a 32-byte map regardless, which could cause it to run off the end
4533 of the memory it had got.
4534
4535 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4536
4537
4538 Version 0.98 22-Oct-97
4539 ----------------------
4540
4541 1. Fixed bug in code for handling temporary memory usage when there are more
4542 back references than supplied space in the ovector. This could cause segfaults.
4543
4544
4545 Version 0.97 21-Oct-97
4546 ----------------------
4547
4548 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4549
4550 2. Optimized negated single characters not to use a bit map.
4551
4552 3. Brought error texts together as macro definitions; clarified some of them;
4553 fixed one that was wrong - it said "range out of order" when it meant "invalid
4554 escape sequence".
4555
4556 4. Changed some char * arguments to const char *.
4557
4558 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4559
4560 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4561 pcretest.
4562
4563
4564 Version 0.96 16-Oct-97
4565 ----------------------
4566
4567 1. Added a simple "pgrep" utility to the distribution.
4568
4569 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4570 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4571 where "ddd" means "one or more decimal digits".
4572
4573 3. Fixed serious bug. If a pattern had a back reference, but the call to
4574 pcre_exec() didn't supply a large enough ovector to record the related
4575 identifying subpattern, the match always failed. PCRE now remembers the number
4576 of the largest back reference, and gets some temporary memory in which to save
4577 the offsets during matching if necessary, in order to ensure that
4578 backreferences always work.
4579
4580 4. Increased the compatibility with Perl in a number of ways:
4581
4582 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4583 to request this handling. The option can be set at compile or exec time.
4584
4585 (b) $ matches before a terminating newline by default; an option
4586 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4587 mode). The option can be set at compile or exec time.
4588
4589 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4590 the same as Perl's. If the decimal number it represents is less than 10
4591 or there aren't that many previous left capturing parentheses, an octal
4592 escape is read. Inside a character class, it's always an octal escape,
4593 even if it is a single digit.
4594
4595 (d) An escaped but undefined alphabetic character is taken as a literal,
4596 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4597 escapes.
4598
4599 (e) {0} is now permitted. (The previous item is removed from the compiled
4600 pattern).
4601
4602 5. Changed all the names of code files so that the basic parts are no longer
4603 than 10 characters, and abolished the teeny "globals.c" file.
4604
4605 6. Changed the handling of character classes; they are now done with a 32-byte
4606 bit map always.
4607
4608 7. Added the -d and /D options to pcretest to make it possible to look at the
4609 internals of compilation without having to recompile pcre.
4610
4611
4612 Version 0.95 23-Sep-97
4613 ----------------------
4614
4615 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4616 \x20 at the start of a run of normal characters. These were being treated as
4617 real characters, instead of the source characters being re-checked.
4618
4619
4620 Version 0.94 18-Sep-97
4621 ----------------------
4622
4623 1. The functions are now thread-safe, with the caveat that the global variables
4624 containing pointers to malloc() and free() or alternative functions are the
4625 same for all threads.
4626
4627 2. Get pcre_study() to generate a bitmap of initial characters for non-
4628 anchored patterns when this is possible, and use it if passed to pcre_exec().
4629
4630
4631 Version 0.93 15-Sep-97
4632 ----------------------
4633
4634 1. /(b)|(:+)/ was computing an incorrect first character.
4635
4636 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4637 but not actually doing anything yet.
4638
4639 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4640 as Perl does (e.g. [-az] or [az-]).
4641
4642 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4643 all possible positions.
4644
4645 5. Split up into different modules to avoid including unneeded functions in a
4646 compiled binary. However, compile and exec are still in one module. The "study"
4647 function is split off.
4648
4649 6. The character tables are now in a separate module whose source is generated
4650 by an auxiliary program - but can then be edited by hand if required. There are
4651 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4652 toupper() in the code.
4653
4654 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4655 make them global. Abolish the function for setting them, as the caller can now
4656 set them directly.
4657
4658
4659 Version 0.92 11-Sep-97
4660 ----------------------
4661
4662 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4663 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4664
4665 2. Caseless matching was not working in character classes if the characters in
4666 the pattern were in upper case.
4667
4668 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4669
4670 4. Make PCRE_ANCHORED public and accept as a compile option.
4671
4672 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4673 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4674 pass them.
4675
4676 6. Give an error if bad option bits passed at compile or run time.
4677
4678 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4679 pcretest to cause it to pass that flag.
4680
4681 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4682 options, and the first character, if set.
4683
4684 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4685
4686
4687 Version 0.91 10-Sep-97
4688 ----------------------
4689
4690 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4691 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4692
4693 2. PCRE was looping on encountering an indefinitely repeated back reference to
4694 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4695 Perl does - treats the match as successful.
4696
4697 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5