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

Diff of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 474 by ph10, Sat Jan 2 16:30:46 2010 UTC revision 926 by ph10, Wed Feb 22 15:01:32 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 8.01 11-Dec-09  Version 8.31
5  ----------------------  -----------------------------
6    
7    1.  Fixing a wrong JIT test case and some compiler warnings.
8    
9    2.  Removed a bashism from the RunTest script.
10    
11    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
12        to unsigned type, result still unsigned" that was given by an MS compiler
13        on encountering the code "-sizeof(xxx)".
14    
15    4.  Partial matching support is added to the JIT compiler.
16    
17    5.  Fixed several bugs concerned with partial matching of items that consist
18        of more than one character:
19    
20        (a) /^(..)\1/ did not partially match "aba" because checking references was
21            done on an "all or nothing" basis. This also applied to repeated
22            references.
23    
24        (b) \R did not give a hard partial match if \r was found at the end of the
25            subject.
26    
27        (c) \X did not give a hard partial match after matching one or more
28            characters at the end of the subject.
29    
30        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
31            a partial match for the string "\r".
32    
33        (e) When newline was set to CRLF, the metacharacter "." did not recognize
34            a partial match for a CR character at the end of the subject string.
35    
36    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
37        running pcretest, the text "(JIT)" added to the output whenever JIT is
38        actually used to run the match.
39    
40    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
41        or /S+[+] with a digit between 1 and 7.
42    
43    8.  OP_NOT now supports any UTF character not just single-byte ones.
44    
45    
46    Version 8.30 04-February-2012
47    -----------------------------
48    
49    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
50        name is defined in ctype.h.
51    
52    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
53        only in quite long subpatterns.
54    
55    3.  Removed the function pcre_info(), which has been obsolete and deprecated
56        since it was replaced by pcre_fullinfo() in February 2000.
57    
58    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
59        match a (*MARK), and the match failed at the start of the subject, a
60        reference to memory before the start of the subject could occur. This bug
61        was introduced by fix 17 of release 8.21.
62    
63    5.  A reference to an unset group with zero minimum repetition was giving
64        totally wrong answers (in non-JavaScript-compatibility mode). For example,
65        /(another)?(\1?)test/ matched against "hello world test". This bug was
66        introduced in release 8.13.
67    
68    6.  Add support for 16-bit character strings (a large amount of work involving
69        many changes and refactorings).
70    
71    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
72        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
73        from a file.
74    
75    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
76        rounding is not applied in this particular case).
77    
78    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
79        if they appear, or are escaped, in patterns.
80    
81    10. Get rid of a number of -Wunused-but-set-variable warnings.
82    
83    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
84        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
85        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
86        also returns the mark "x". This bug applied to capturing parentheses,
87        non-capturing parentheses, and atomic parentheses. It also applied to some
88        assertions.
89    
90    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
91        information out of configure.ac instead of relying on pcre.h.generic, which
92        is not stored in the repository.
93    
94    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
95        -lreadline.
96    
97    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
98    
99    15. Applied Graycode's patch to put the top-level frame on the stack rather
100        than the heap when not using the stack for recursion. This gives a
101        performance improvement in many cases when recursion is not deep.
102    
103    16. Experimental code added to "pcretest -C" to output the stack frame size.
104    
105    
106    Version 8.21 12-Dec-2011
107    ------------------------
108    
109    1.  Updating the JIT compiler.
110    
111    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
112        are added as well.
113    
114    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
115        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
116        calling _pcre_jit_exec. Some extra comments are added.
117    
118    4.  (*MARK) settings inside atomic groups that do not contain any capturing
119        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
120        was introduced by change 18 for 8.20.
121    
122    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
123        ECMA-262 standard.
124    
125    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
126        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
127        This bug was probably introduced by change 9 of 8.13.
128    
129    7.  While fixing 6 above, I noticed that a number of other items were being
130        incorrectly rejected as "not fixed length". This arose partly because newer
131        opcodes had not been added to the fixed-length checking code. I have (a)
132        corrected the bug and added tests for these items, and (b) arranged for an
133        error to occur if an unknown opcode is encountered while checking for fixed
134        length instead of just assuming "not fixed length". The items that were
135        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
136        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
137        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
138    
139    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
140        being incorrectly compiled and would have given unpredicatble results.
141    
142    9.  A possessively repeated subpattern with minimum repeat count greater than
143        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
144        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
145        the first (A) could occur when it should not.
146    
147    10. Add a cast and remove a redundant test from the code.
148    
149    11. JIT should use pcre_malloc/pcre_free for allocation.
150    
151    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
152        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
153        is anything other than /usr, -L is still shown).
154    
155    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
156    
157    14. Perl does not support \N without a following name in a [] class; PCRE now
158        also gives an error.
159    
160    15. If a forward reference was repeated with an upper limit of around 2000,
161        it caused the error "internal error: overran compiling workspace". The
162        maximum number of forward references (including repeats) was limited by the
163        internal workspace, and dependent on the LINK_SIZE. The code has been
164        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
165        the default depends on LINK_SIZE. There is a new upper limit (for safety)
166        of around 200,000 forward references. While doing this, I also speeded up
167        the filling in of repeated forward references.
168    
169    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
170        incorrectly expecting the subject to contain another "a" after the start.
171    
172    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
173        in the match, the SKIP should be ignored. This was not happening; instead
174        the SKIP was being treated as NOMATCH. For patterns such as
175        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
176        tested.
177    
178    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
179        now much more compatible with Perl, in particular in cases where the result
180        is a non-match for a non-anchored pattern. For example, if
181        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
182        "m", where previously it did not return a name. A side effect of this
183        change is that for partial matches, the last encountered mark name is
184        returned, as for non matches. A number of tests that were previously not
185        Perl-compatible have been moved into the Perl-compatible test files. The
186        refactoring has had the pleasing side effect of removing one argument from
187        the match() function, thus reducing its stack requirements.
188    
189    19. If the /S+ option was used in pcretest to study a pattern using JIT,
190        subsequent uses of /S (without +) incorrectly behaved like /S+.
191    
192    21. Retrieve executable code size support for the JIT compiler and fixing
193        some warnings.
194    
195    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
196        not work when the shorter character appeared right at the end of the
197        subject string.
198    
199    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
200        systems.
201    
202    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
203        output it when the /M option is used in pcretest.
204    
205    25. The CheckMan script was not being included in the distribution. Also, added
206        an explicit "perl" to run Perl scripts from the PrepareRelease script
207        because this is reportedly needed in Windows.
208    
209    26. If study data was being save in a file and studying had not found a set of
210        "starts with" bytes for the pattern, the data written to the file (though
211        never used) was taken from uninitialized memory and so caused valgrind to
212        complain.
213    
214    27. Updated RunTest.bat as provided by Sheri Pierce.
215    
216    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
217    
218    29. Computation of memory usage for the table of capturing group names was
219        giving an unnecessarily large value.
220    
221    
222    Version 8.20 21-Oct-2011
223    ------------------------
224    
225    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
226        a POSIX class. After further experiments with Perl, which convinced me that
227        Perl has bugs and confusions, a closing square bracket is no longer allowed
228        in a POSIX name. This bug also affected patterns with classes that started
229        with full stops.
230    
231    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
232        captured substring, but while checking the failing first alternative,
233        substring 1 is temporarily captured. If the output vector supplied to
234        pcre_exec() was not big enough for this capture, the yield of the function
235        was still zero ("insufficient space for captured substrings"). This cannot
236        be totally fixed without adding another stack variable, which seems a lot
237        of expense for a edge case. However, I have improved the situation in cases
238        such as /(a)(b)x|abc/ matched against "abc", where the return code
239        indicates that fewer than the maximum number of slots in the ovector have
240        been set.
241    
242    3.  Related to (2) above: when there are more back references in a pattern than
243        slots in the output vector, pcre_exec() uses temporary memory during
244        matching, and copies in the captures as far as possible afterwards. It was
245        using the entire output vector, but this conflicts with the specification
246        that only 2/3 is used for passing back captured substrings. Now it uses
247        only the first 2/3, for compatibility. This is, of course, another edge
248        case.
249    
250    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
251        main code base, and can be used by building with --enable-jit. When this is
252        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
253        runtime --no-jit option is given.
254    
255    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
256        ovector, the return from the function was zero, implying that there were
257        other matches that did not fit. The correct "exactly full" value is now
258        returned.
259    
260    6.  If a subpattern that was called recursively or as a subroutine contained
261        (*PRUNE) or any other control that caused it to give a non-standard return,
262        invalid errors such as "Error -26 (nested recursion at the same subject
263        position)" or even infinite loops could occur.
264    
265    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
266        computing the minimum length on reaching *ACCEPT, and so ended up with the
267        wrong value of 1 rather than 0. Further investigation indicates that
268        computing a minimum subject length in the presence of *ACCEPT is difficult
269        (think back references, subroutine calls), and so I have changed the code
270        so that no minimum is registered for a pattern that contains *ACCEPT.
271    
272    8.  If (*THEN) was present in the first (true) branch of a conditional group,
273        it was not handled as intended. [But see 16 below.]
274    
275    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
276        Sheri Pierce.
277    
278    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
279        the first byte in a match must be "a".
280    
281    11. Change 17 for 8.13 increased the recursion depth for patterns like
282        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
283        pattern contains any instances of (*THEN). If it does not, the old
284        optimizations are restored. It would be nice to do this on a per-group
285        basis, but at the moment that is not feasible.
286    
287    12. In some environments, the output of pcretest -C is CRLF terminated. This
288        broke RunTest's code that checks for the link size. A single white space
289        character after the value is now allowed for.
290    
291    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
292        For "fr", it uses the Windows-specific input and output files.
293    
294    14. If (*THEN) appeared in a group that was called recursively or as a
295        subroutine, it did not work as intended. [But see next item.]
296    
297    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
298        pattern fragments (but not containing any | characters). If A and B are
299        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
300        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
301        D. In other words, Perl considers parentheses that do not contain any |
302        characters to be part of a surrounding alternative, whereas PCRE was
303        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
304        differently. PCRE now behaves in the same way as Perl, except in the case
305        of subroutine/recursion calls such as (?1) which have in any case always
306        been different (but PCRE had them first :-).
307    
308    16. Related to 15 above: Perl does not treat the | in a conditional group as
309        creating alternatives. Such a group is treated in the same way as an
310        ordinary group without any | characters when processing (*THEN). PCRE has
311        been changed to match Perl's behaviour.
312    
313    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
314        RunGrepTest script failed.
315    
316    18. Change 22 for version 13 caused atomic groups to use more stack. This is
317        inevitable for groups that contain captures, but it can lead to a lot of
318        stack use in large patterns. The old behaviour has been restored for atomic
319        groups that do not contain any capturing parentheses.
320    
321    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
322        suppress the check for a minimum subject length at run time. (If it was
323        given to pcre_exec() or pcre_dfa_exec() it did work.)
324    
325    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
326        fail to work when decoding hex characters in data strings in EBCDIC
327        environments.
328    
329    21. It appears that in at least one Mac OS environment, the isxdigit() function
330        is implemented as a macro that evaluates to its argument more than once,
331        contravening the C 90 Standard (I haven't checked a later standard). There
332        was an instance in pcretest which caused it to go wrong when processing
333        \x{...} escapes in subject strings. The has been rewritten to avoid using
334        things like p++ in the argument of isxdigit().
335    
336    
337    Version 8.13 16-Aug-2011
338    ------------------------
339    
340    1.  The Unicode data tables have been updated to Unicode 6.0.0.
341    
342    2.  Two minor typos in pcre_internal.h have been fixed.
343    
344    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
345        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
346        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
347    
348    4.  There were a number of related bugs in the code for matching backrefences
349        caselessly in UTF-8 mode when codes for the characters concerned were
350        different numbers of bytes. For example, U+023A and U+2C65 are an upper
351        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
352        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
353        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
354        2-byte code at the end of the subject (it thought there wasn't enough data
355        left).
356    
357    5.  Comprehensive information about what went wrong is now returned by
358        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
359        as the output vector has at least 2 elements. The offset of the start of
360        the failing character and a reason code are placed in the vector.
361    
362    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
363        now returned is for the first byte of the failing character, instead of the
364        last byte inspected. This is an incompatible change, but I hope it is small
365        enough not to be a problem. It makes the returned offset consistent with
366        pcre_exec() and pcre_dfa_exec().
367    
368    7.  pcretest now gives a text phrase as well as the error number when
369        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
370        failure, the offset and reason code are output.
371    
372    8.  When \R was used with a maximizing quantifier it failed to skip backwards
373        over a \r\n pair if the subsequent match failed. Instead, it just skipped
374        back over a single character (\n). This seems wrong (because it treated the
375        two characters as a single entity when going forwards), conflicts with the
376        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
377        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
378        has been changed.
379    
380    9.  Some internal refactoring has changed the processing so that the handling
381        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
382        time (the PCRE_DOTALL option was changed this way some time ago: version
383        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
384        which was always a bit of a fudge. It also means that there is one less
385        argument for the match() function, which reduces its stack requirements
386        slightly. This change also fixes an incompatibility with Perl: the pattern
387        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
388    
389    10. More internal refactoring has drastically reduced the number of recursive
390        calls to match() for possessively repeated groups such as (abc)++ when
391        using pcre_exec().
392    
393    11. While implementing 10, a number of bugs in the handling of groups were
394        discovered and fixed:
395    
396        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
397        (a|)*(?1) gave a compile-time internal error.
398        ((a|)+)+  did not notice that the outer group could match an empty string.
399        (^a|^)+   was not marked as anchored.
400        (.*a|.*)+ was not marked as matching at start or after a newline.
401    
402    12. Yet more internal refactoring has removed another argument from the match()
403        function. Special calls to this function are now indicated by setting a
404        value in a variable in the "match data" data block.
405    
406    13. Be more explicit in pcre_study() instead of relying on "default" for
407        opcodes that mean there is no starting character; this means that when new
408        ones are added and accidentally left out of pcre_study(), testing should
409        pick them up.
410    
411    14. The -s option of pcretest has been documented for ages as being an old
412        synonym of -m (show memory usage). I have changed it to mean "force study
413        for every regex", that is, assume /S for every regex. This is similar to -i
414        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
415        using it. It makes it easier to run collections of tests with and without
416        study enabled, and thereby test pcre_study() more easily. All the standard
417        tests are now run with and without -s (but some patterns can be marked as
418        "never study" - see 20 below).
419    
420    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
421        restoration of the capturing data to the outer values was not happening
422        correctly.
423    
424    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
425        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
426        pattern had matched an empty string, and so incorrectly returned a no
427        match.
428    
429    17. There was optimizing code for the last branch of non-capturing parentheses,
430        and also for the obeyed branch of a conditional subexpression, which used
431        tail recursion to cut down on stack usage. Unfortunately, now that there is
432        the possibility of (*THEN) occurring in these branches, tail recursion is
433        no longer possible because the return has to be checked for (*THEN). These
434        two optimizations have therefore been removed. [But see 8.20/11 above.]
435    
436    18. If a pattern containing \R was studied, it was assumed that \R always
437        matched two bytes, thus causing the minimum subject length to be
438        incorrectly computed because \R can also match just one byte.
439    
440    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
441        was incorrectly computed.
442    
443    20. If /S is present twice on a test pattern in pcretest input, it now
444        *disables* studying, thereby overriding the use of -s on the command line
445        (see 14 above). This is necessary for one or two tests to keep the output
446        identical in both cases.
447    
448    21. When (*ACCEPT) was used in an assertion that matched an empty string and
449        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
450    
451    22. When an atomic group that contained a capturing parenthesis was
452        successfully matched, but the branch in which it appeared failed, the
453        capturing was not being forgotten if a higher numbered group was later
454        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
455        group 1 to "a", when in fact it should be unset. This applied to multi-
456        branched capturing and non-capturing groups, repeated or not, and also to
457        positive assertions (capturing in negative assertions does not happen
458        in PCRE) and also to nested atomic groups.
459    
460    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
461        subject after a captured substring, to make it easier to tell which of a
462        number of identical substrings has been captured.
463    
464    24. The way atomic groups are processed by pcre_exec() has been changed so that
465        if they are repeated, backtracking one repetition now resets captured
466        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
467        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
468        "aaa". Previously, it would have been "a". As part of this code
469        refactoring, the way recursive calls are handled has also been changed.
470    
471    25. If an assertion condition captured any substrings, they were not passed
472        back unless some other capturing happened later. For example, if
473        (?(?=(a))a) was matched against "a", no capturing was returned.
474    
475    26. When studying a pattern that contained subroutine calls or assertions,
476        the code for finding the minimum length of a possible match was handling
477        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
478        group 1 called group 2 while simultaneously a separate group 2 called group
479        1). A stack overflow occurred in this case. I have fixed this by limiting
480        the recursion depth to 10.
481    
482    27. Updated RunTest.bat in the distribution to the version supplied by Tom
483        Fortmann. This supports explicit test numbers on the command line, and has
484        argument validation and error reporting.
485    
486    28. An instance of \X with an unlimited repeat could fail if at any point the
487        first character it looked at was a mark character.
488    
489    29. Some minor code refactoring concerning Unicode properties and scripts
490        should reduce the stack requirement of match() slightly.
491    
492    30. Added the '=' option to pcretest to check the setting of unused capturing
493        slots at the end of the pattern, which are documented as being -1, but are
494        not included in the return count.
495    
496    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
497        compiled something random. Now it gives a compile-time error (as does
498        Perl).
499    
500    32. A *MARK encountered during the processing of a positive assertion is now
501        recorded and passed back (compatible with Perl).
502    
503    33. If --only-matching or --colour was set on a pcregrep call whose pattern
504        had alternative anchored branches, the search for a second match in a line
505        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
506        matched the line "0102" twice. The same bug affected patterns that started
507        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
508        twice.
509    
510    34. Previously, PCRE did not allow quantification of assertions. However, Perl
511        does, and because of capturing effects, quantifying parenthesized
512        assertions may at times be useful. Quantifiers are now allowed for
513        parenthesized assertions.
514    
515    35. A minor code tidy in pcre_compile() when checking options for \R usage.
516    
517    36. \g was being checked for fancy things in a character class, when it should
518        just be a literal "g".
519    
520    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
521        appearance of a nested POSIX class supersedes an apparent external class.
522        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
523        unescaped square brackets may also appear as part of class names. For
524        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
525        more like Perl. (But see 8.20/1 above.)
526    
527    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
528        was because it thought it was \N{name}, which is not supported).
529    
530    39. Add minix to OS list not supporting the -S option in pcretest.
531    
532    40. PCRE tries to detect cases of infinite recursion at compile time, but it
533        cannot analyze patterns in sufficient detail to catch mutual recursions
534        such as ((?1))((?2)). There is now a runtime test that gives an error if a
535        subgroup is called recursively as a subpattern for a second time at the
536        same position in the subject string. In previous releases this might have
537        been caught by the recursion limit, or it might have run out of stack.
538    
539    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
540        happen only once. PCRE was, however incorrectly giving a compile time error
541        "recursive call could loop indefinitely" because it cannot analyze the
542        pattern in sufficient detail. The compile time test no longer happens when
543        PCRE is compiling a conditional subpattern, but actual runaway loops are
544        now caught at runtime (see 40 above).
545    
546    42. It seems that Perl allows any characters other than a closing parenthesis
547        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
548        has been changed to be the same.
549    
550    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
551        as not to get warnings when autogen.sh is called. Also changed
552        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
553    
554    44. To help people who use pcregrep to scan files containing exceedingly long
555        lines, the following changes have been made:
556    
557        (a) The default value of the buffer size parameter has been increased from
558            8K to 20K. (The actual buffer used is three times this size.)
559    
560        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
561            PCRE is built.
562    
563        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
564            to be set at run time.
565    
566        (d) Numerical values in pcregrep options can be followed by K or M, for
567            example --buffer-size=50K.
568    
569        (e) If a line being scanned overflows pcregrep's buffer, an error is now
570            given and the return code is set to 2.
571    
572    45. Add a pointer to the latest mark to the callout data block.
573    
574    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
575        partial match of an empty string instead of no match. This was specific to
576        the use of ".".
577    
578    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
579        complete match instead of a partial match. This bug was dependent on both
580        the PCRE_UTF8 and PCRE_DOTALL options being set.
581    
582    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
583        starting byte set, because \b was not being ignored.
584    
585    
586    Version 8.12 15-Jan-2011
587    ------------------------
588    
589    1.  Fixed some typos in the markup of the man pages, and wrote a script that
590        checks for such things as part of the documentation building process.
591    
592    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
593        --match-limit and --recursion-limit options (added for 8.11). In
594        particular, this made one of the standard tests fail. (The integer value
595        went into the wrong half of a long int.)
596    
597    3.  If the --colour option was given to pcregrep with -v (invert match), it
598        did strange things, either producing crazy output, or crashing. It should,
599        of course, ignore a request for colour when reporting lines that do not
600        match.
601    
602    4.  Another pcregrep bug caused similar problems if --colour was specified with
603        -M (multiline) and the pattern match finished with a line ending.
604    
605    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
606        matched in multiline mode, the following line was shown as part of the
607        match. This seems wrong, so I have changed it.
608    
609    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
610        the check for further matches in the same line (so they could be coloured)
611        to overrun the end of the current line. If another match was found, it was
612        incorrectly shown (and then shown again when found in the next line).
613    
614    7.  If pcregrep was compiled under Windows, there was a reference to the
615        function pcregrep_exit() before it was defined. I am assuming this was
616        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
617        reported by a user. I've moved the definition above the reference.
618    
619    
620    Version 8.11 10-Dec-2010
621    ------------------------
622    
623    1.  (*THEN) was not working properly if there were untried alternatives prior
624        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
625        backtracked to try for "b" instead of moving to the next alternative branch
626        at the same level (in this case, to look for "c"). The Perl documentation
627        is clear that when (*THEN) is backtracked onto, it goes to the "next
628        alternative in the innermost enclosing group".
629    
630    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
631        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
632        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
633        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
634        (*THEN).
635    
636    3.  If \s appeared in a character class, it removed the VT character from
637        the class, even if it had been included by some previous item, for example
638        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
639        of \s, but is part of the POSIX "space" class.)
640    
641    4.  A partial match never returns an empty string (because you can always
642        match an empty string at the end of the subject); however the checking for
643        an empty string was starting at the "start of match" point. This has been
644        changed to the "earliest inspected character" point, because the returned
645        data for a partial match starts at this character. This means that, for
646        example, /(?<=abc)def/ gives a partial match for the subject "abc"
647        (previously it gave "no match").
648    
649    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
650        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
651        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
652        has an implication that the given string is incomplete (because a partial
653        match is preferred over a full match). For this reason, these items now
654        give a partial match in this situation. [Aside: previously, the one case
655        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
656        match rather than a full match, which was wrong by the old rules, but is
657        now correct.]
658    
659    6.  There was a bug in the handling of #-introduced comments, recognized when
660        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
661        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
662        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
663        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
664        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
665        places in pcre_compile().
666    
667    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
668        comments when looking ahead for named forward references to subpatterns,
669        the only newline sequence it recognized was NL. It now handles newlines
670        according to the set newline convention.
671    
672    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
673        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
674        cater for a lack of strerror(). These oversights have been fixed.
675    
676    9.  Added --match-limit and --recursion-limit to pcregrep.
677    
678    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
679    
680    11. When the -o option was used, pcregrep was setting a return code of 1, even
681        when matches were found, and --line-buffered was not being honoured.
682    
683    12. Added an optional parentheses number to the -o and --only-matching options
684        of pcregrep.
685    
686    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
687        can match an empty string. The code to do it in pcretest and pcredemo
688        needed fixing:
689    
690        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
691            only one byte after an empty string match just before CRLF (this case
692            just got forgotten; "any" and "anycrlf" were OK).
693    
694        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
695            mode when an empty string match preceded an ASCII character followed by
696            a non-ASCII character. (The code for advancing by one character rather
697            than one byte was nonsense.)
698    
699        (c) The pcredemo.c sample program did not have any code at all to handle
700            the cases when CRLF is a valid newline sequence.
701    
702    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
703        as a starting offset was within the subject string. There is now a new
704        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
705        negative or greater than the length of the string. In order to test this,
706        pcretest is extended to allow the setting of negative starting offsets.
707    
708    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
709        starting offset points to the beginning of a UTF-8 character was
710        unnecessarily clumsy. I tidied it up.
711    
712    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
713        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
714    
715    17. Nobody had reported that the --include_dir option, which was added in
716        release 7.7 should have been called --include-dir (hyphen, not underscore)
717        for compatibility with GNU grep. I have changed it to --include-dir, but
718        left --include_dir as an undocumented synonym, and the same for
719        --exclude-dir, though that is not available in GNU grep, at least as of
720        release 2.5.4.
721    
722    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
723        characters from a string of bytes) have been redefined so as not to use
724        loops, in order to improve performance in some environments. At the same
725        time, I abstracted some of the common code into auxiliary macros to save
726        repetition (this should not affect the compiled code).
727    
728    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
729        compile-time error is now given if \c is not followed by an ASCII
730        character, that is, a byte less than 128. (In EBCDIC mode, the code is
731        different, and any byte value is allowed.)
732    
733    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
734        START_OPTIMIZE option, which is now allowed at compile time - but just
735        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
736        to pcregrep and other applications that have no direct access to PCRE
737        options. The new /Y option in pcretest sets this option when calling
738        pcre_compile().
739    
740    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
741        back references. Groups containing recursive back references were forced to
742        be atomic by that change, but in the case of named groups, the amount of
743        memory required was incorrectly computed, leading to "Failed: internal
744        error: code overflow". This has been fixed.
745    
746    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
747        pcretest.c, to avoid build problems in some Borland environments.
748    
749    
750    Version 8.10 25-Jun-2010
751    ------------------------
752    
753    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
754        THEN.
755    
756    2.  (*ACCEPT) was not working when inside an atomic group.
757    
758    3.  Inside a character class, \B is treated as a literal by default, but
759        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
760        causes the error). The code is unchanged, but I tidied the documentation.
761    
762    4.  Inside a character class, PCRE always treated \R and \X as literals,
763        whereas Perl faults them if its -w option is set. I have changed PCRE so
764        that it faults them when PCRE_EXTRA is set.
765    
766    5.  Added support for \N, which always matches any character other than
767        newline. (It is the same as "." when PCRE_DOTALL is not set.)
768    
769    6.  When compiling pcregrep with newer versions of gcc which may have
770        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
771        declared with attribute warn_unused_result" were given. Just casting the
772        result to (void) does not stop the warnings; a more elaborate fudge is
773        needed. I've used a macro to implement this.
774    
775    7.  Minor change to pcretest.c to avoid a compiler warning.
776    
777    8.  Added four artifical Unicode properties to help with an option to make
778        \s etc use properties (see next item). The new properties are: Xan
779        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
780    
781    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
782        use Unicode properties. (*UCP) at the start of a pattern can be used to set
783        this option. Modified pcretest to add /W to test this facility. Added
784        REG_UCP to make it available via the POSIX interface.
785    
786    10. Added --line-buffered to pcregrep.
787    
788    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
789        studied, and the match started with a letter with a code point greater than
790        127 whose first byte was different to the first byte of the other case of
791        the letter, the other case of this starting letter was not recognized
792        (#976).
793    
794    12. If a pattern that was studied started with a repeated Unicode property
795        test, for example, \p{Nd}+, there was the theoretical possibility of
796        setting up an incorrect bitmap of starting bytes, but fortunately it could
797        not have actually happened in practice until change 8 above was made (it
798        added property types that matched character-matching opcodes).
799    
800    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
801        possible starting bytes for non-anchored patterns.
802    
803    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
804        \R, and also a number of cases that involve Unicode properties, both
805        explicit and implicit when PCRE_UCP is set.
806    
807    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
808        input, it could crash or give wrong results if characters with values
809        greater than 0xc0 were present in the subject string. (Detail: it assumed
810        UTF-8 input when processing these items.)
811    
812    16. Added a lot of (int) casts to avoid compiler warnings in systems where
813        size_t is 64-bit (#991).
814    
815    17. Added a check for running out of memory when PCRE is compiled with
816        --disable-stack-for-recursion (#990).
817    
818    18. If the last data line in a file for pcretest does not have a newline on
819        the end, a newline was missing in the output.
820    
821    19. The default pcre_chartables.c file recognizes only ASCII characters (values
822        less than 128) in its various bitmaps. However, there is a facility for
823        generating tables according to the current locale when PCRE is compiled. It
824        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
825        characters, are recognized by isspace() and therefore were getting set in
826        these tables, and indeed these tables seem to approximate to ISO 8859. This
827        caused a problem in UTF-8 mode when pcre_study() was used to create a list
828        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
829        which of course cannot start UTF-8 characters. I have changed the code so
830        that only real ASCII characters (less than 128) and the correct starting
831        bytes for UTF-8 encodings are set for characters greater than 127 when in
832        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
833        altogether.)
834    
835    20. Added the /T option to pcretest so as to be able to run tests with non-
836        standard character tables, thus making it possible to include the tests
837        used for 19 above in the standard set of tests.
838    
839    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
840        reference to a subpattern the other side of a comment that contains an
841        opening parenthesis caused either an internal compiling error, or a
842        reference to the wrong subpattern.
843    
844    
845    Version 8.02 19-Mar-2010
846    ------------------------
847    
848    1.  The Unicode data tables have been updated to Unicode 5.2.0.
849    
850    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
851        configured.
852    
853    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
854        original author of that file, following a query about its status.
855    
856    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
857        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
858    
859    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
860        quantifier applied to a forward-referencing subroutine call, could compile
861        incorrect code or give the error "internal error: previously-checked
862        referenced subpattern not found".
863    
864    6.  Both MS Visual Studio and Symbian OS have problems with initializing
865        variables to point to external functions. For these systems, therefore,
866        pcre_malloc etc. are now initialized to local functions that call the
867        relevant global functions.
868    
869    7.  There were two entries missing in the vectors called coptable and poptable
870        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
871        I've fixed the data, and added a kludgy way of testing at compile time that
872        the lengths are correct (equal to the number of opcodes).
873    
874    8.  Following on from 7, I added a similar kludge to check the length of the
875        eint vector in pcreposix.c.
876    
877    9.  Error texts for pcre_compile() are held as one long string to avoid too
878        much relocation at load time. To find a text, the string is searched,
879        counting zeros. There was no check for running off the end of the string,
880        which could happen if a new error number was added without updating the
881        string.
882    
883    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
884    
885    11. \K was not working if it appeared in an atomic group or in a group that
886        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
887        \K is "not well defined" if used in an assertion. PCRE now accepts it if
888        the assertion is positive, but not if it is negative.
889    
890    12. Change 11 fortuitously reduced the size of the stack frame used in the
891        "match()" function of pcre_exec.c by one pointer. Forthcoming
892        implementation of support for (*MARK) will need an extra pointer on the
893        stack; I have reserved it now, so that the stack frame size does not
894        decrease.
895    
896    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
897        item in branch that calls a recursion is a subroutine call - as in the
898        second branch in the above example - was incorrectly given the compile-
899        time error "recursive call could loop indefinitely" because pcre_compile()
900        was not correctly checking the subroutine for matching a non-empty string.
901    
902    14. The checks for overrunning compiling workspace could trigger after an
903        overrun had occurred. This is a "should never occur" error, but it can be
904        triggered by pathological patterns such as hundreds of nested parentheses.
905        The checks now trigger 100 bytes before the end of the workspace.
906    
907    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
908    
909    
910    Version 8.01 19-Jan-2010
911    ------------------------
912    
913  1.  If a pattern contained a conditional subpattern with only one branch (in  1.  If a pattern contained a conditional subpattern with only one branch (in
914      particular, this includes all (DEFINE) patterns), a call to pcre_study()      particular, this includes all (*DEFINE) patterns), a call to pcre_study()
915      computed the wrong minimum data length (which is of course zero for such      computed the wrong minimum data length (which is of course zero for such
916      subpatterns).      subpatterns). This could cause incorrect "no match" results.
917    
918  2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of  2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
919      the pattern is reset in the first branch, pcre_compile() failed with      the pattern is reset in the first branch, pcre_compile() failed with
# Line 26  Version 8.01 11-Dec-09 Line 932  Version 8.01 11-Dec-09
932      assertion subpattern, including such a pattern used as a condition,      assertion subpattern, including such a pattern used as a condition,
933      unpredictable results occurred, instead of the error return      unpredictable results occurred, instead of the error return
934      PCRE_ERROR_DFA_UITEM.      PCRE_ERROR_DFA_UITEM.
935    
936  5.  The C++ GlobalReplace function was not working like Perl for the special  5.  The C++ GlobalReplace function was not working like Perl for the special
937      situation when an empty string is matched. It now does the fancy magic      situation when an empty string is matched. It now does the fancy magic
938      stuff that is necessary.      stuff that is necessary.
939    
940    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
941        removed. (These were left over from very, very early versions of PCRE.)
942    
943    7.  Some cosmetic changes to the code to make life easier when compiling it
944        as part of something else:
945    
946        (a) Change DEBUG to PCRE_DEBUG.
947    
948        (b) In pcre_compile(), rename the member of the "branch_chain" structure
949            called "current" as "current_branch", to prevent a collision with the
950            Linux macro when compiled as a kernel module.
951    
952        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
953            prevent a collision with the Linux macro when compiled as a kernel
954            module.
955    
956    8.  In pcre_compile() there are some checks for integer overflows that used to
957        cast potentially large values to (double). This has been changed to that
958        when building, a check for int64_t is made, and if it is found, it is used
959        instead, thus avoiding the use of floating point arithmetic. (There is no
960        other use of FP in PCRE.) If int64_t is not found, the fallback is to
961        double.
962    
963    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
964        2005 (difference between two addresses compared to an unsigned value).
965    
966    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
967        custom one, because of the following reported problem in Windows:
968    
969          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
970              under Win32.
971          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
972              therefore missing the function definition.
973          - The compiler thus generates a "C" signature for the test function.
974          - The linker fails to find the "C" function.
975          - PCRE fails to configure if asked to do so against libbz2.
976    
977    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
978        messages were output:
979    
980          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
981          rerunning libtoolize, to keep the correct libtool macros in-tree.
982          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
983    
984        I have done both of these things.
985    
986    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
987        most of the time, it *can* run out if it is given a pattern that contains a
988        runaway infinite recursion. I updated the discussion in the pcrestack man
989        page.
990    
991    13. Now that we have gone to the x.xx style of version numbers, the minor
992        version may start with zero. Using 08 or 09 is a bad idea because users
993        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
994        interpreted as invalid octal numbers. I've updated the previous comment in
995        configure.ac, and also added a check that gives an error if 08 or 09 are
996        used.
997    
998    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
999        causing partial matching to fail when the end of the subject matched \W
1000        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1001    
1002    15. There were some discrepancies between the declarations in pcre_internal.h
1003        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1004        their definitions. The declarations used "const uschar *" and the
1005        definitions used USPTR. Even though USPTR is normally defined as "const
1006        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1007        reported that: "This difference in casting confuses some C++ compilers, for
1008        example, SunCC recognizes above declarations as different functions and
1009        generates broken code for hbpcre." I have changed the declarations to use
1010        USPTR.
1011    
1012    16. GNU libtool is named differently on some systems. The autogen.sh script now
1013        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1014        (FreeBSD).
1015    
1016    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1017        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1018        comment: "Figure out how to create a longlong from a string: strtoll and
1019        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1020        instance, but it only takes 2 args instead of 3!"
1021    
1022    18. A subtle bug concerned with back references has been fixed by a change of
1023        specification, with a corresponding code fix. A pattern such as
1024        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1025        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1026        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1027        same bug. Such groups have to be quantified to be useful, or contained
1028        inside another quantified group. (If there's no repetition, the reference
1029        can never match.) The problem arises because, having left the group and
1030        moved on to the rest of the pattern, a later failure that backtracks into
1031        the group uses the captured value from the final iteration of the group
1032        rather than the correct earlier one. I have fixed this in PCRE by forcing
1033        any group that contains a reference to itself to be an atomic group; that
1034        is, there cannot be any backtracking into it once it has completed. This is
1035        similar to recursive and subroutine calls.
1036    
1037    
1038  Version 8.00 19-Oct-09  Version 8.00 19-Oct-09

Legend:
Removed from v.474  
changed lines
  Added in v.926

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12