ViewVC logotype

Diff of /code/trunk/ChangeLog

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

revision 594 by ph10, Sun May 1 08:22:12 2011 UTC revision 741 by zherczeg, Mon Oct 31 09:31:46 2011 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
4  Version 8.13 30-Apr-2011  Version 8.21
5    ------------
7    1.  Updating the JIT compiler.
9    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
10        are added as well.
13    Version 8.20 21-Oct-2011
14    ------------------------
16    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
17        a POSIX class. After further experiments with Perl, which convinced me that
18        Perl has bugs and confusions, a closing square bracket is no longer allowed
19        in a POSIX name. This bug also affected patterns with classes that started
20        with full stops.
22    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
23        captured substring, but while checking the failing first alternative,
24        substring 1 is temporarily captured. If the output vector supplied to
25        pcre_exec() was not big enough for this capture, the yield of the function
26        was still zero ("insufficient space for captured substrings"). This cannot
27        be totally fixed without adding another stack variable, which seems a lot
28        of expense for a edge case. However, I have improved the situation in cases
29        such as /(a)(b)x|abc/ matched against "abc", where the return code
30        indicates that fewer than the maximum number of slots in the ovector have
31        been set.
33    3.  Related to (2) above: when there are more back references in a pattern than
34        slots in the output vector, pcre_exec() uses temporary memory during
35        matching, and copies in the captures as far as possible afterwards. It was
36        using the entire output vector, but this conflicts with the specification
37        that only 2/3 is used for passing back captured substrings. Now it uses
38        only the first 2/3, for compatibility. This is, of course, another edge
39        case.
41    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
42        main code base, and can be used by building with --enable-jit. When this is
43        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
44        runtime --no-jit option is given.
46    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
47        ovector, the return from the function was zero, implying that there were
48        other matches that did not fit. The correct "exactly full" value is now
49        returned.
51    6.  If a subpattern that was called recursively or as a subroutine contained
52        (*PRUNE) or any other control that caused it to give a non-standard return,
53        invalid errors such as "Error -26 (nested recursion at the same subject
54        position)" or even infinite loops could occur.
56    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
57        computing the minimum length on reaching *ACCEPT, and so ended up with the
58        wrong value of 1 rather than 0. Further investigation indicates that
59        computing a minimum subject length in the presence of *ACCEPT is difficult
60        (think back references, subroutine calls), and so I have changed the code
61        so that no minimum is registered for a pattern that contains *ACCEPT.
63    8.  If (*THEN) was present in the first (true) branch of a conditional group,
64        it was not handled as intended. [But see 16 below.]
66    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
67        Sheri Pierce.
69    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
70        the first byte in a match must be "a".
72    11. Change 17 for 8.13 increased the recursion depth for patterns like
73        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
74        pattern contains any instances of (*THEN). If it does not, the old
75        optimizations are restored. It would be nice to do this on a per-group
76        basis, but at the moment that is not feasible.
78    12. In some environments, the output of pcretest -C is CRLF terminated. This
79        broke RunTest's code that checks for the link size. A single white space
80        character after the value is now allowed for.
82    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
83        For "fr", it uses the Windows-specific input and output files.
85    14. If (*THEN) appeared in a group that was called recursively or as a
86        subroutine, it did not work as intended. [But see next item.]
88    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
89        pattern fragments (but not containing any | characters). If A and B are
90        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
91        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
92        D. In other words, Perl considers parentheses that do not contain any |
93        characters to be part of a surrounding alternative, whereas PCRE was
94        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
95        differently. PCRE now behaves in the same way as Perl, except in the case
96        of subroutine/recursion calls such as (?1) which have in any case always
97        been different (but PCRE had them first :-).
99    16. Related to 15 above: Perl does not treat the | in a conditional group as
100        creating alternatives. Such a group is treated in the same way as an
101        ordinary group without any | characters when processing (*THEN). PCRE has
102        been changed to match Perl's behaviour.
104    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
105        RunGrepTest script failed.
107    18. Change 22 for version 13 caused atomic groups to use more stack. This is
108        inevitable for groups that contain captures, but it can lead to a lot of
109        stack use in large patterns. The old behaviour has been restored for atomic
110        groups that do not contain any capturing parentheses.
112    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
113        suppress the check for a minimum subject length at run time. (If it was
114        given to pcre_exec() or pcre_dfa_exec() it did work.)
116    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
117        fail to work when decoding hex characters in data strings in EBCDIC
118        environments.
120    21. It appears that in at least one Mac OS environment, the isxdigit() function
121        is implemented as a macro that evaluates to its argument more than once,
122        contravening the C 90 Standard (I haven't checked a later standard). There
123        was an instance in pcretest which caused it to go wrong when processing
124        \x{...} escapes in subject strings. The has been rewritten to avoid using
125        things like p++ in the argument of isxdigit().
128    Version 8.13 16-Aug-2011
129  ------------------------  ------------------------
131  1.  The Unicode data tables have been updated to Unicode 6.0.0.  1.  The Unicode data tables have been updated to Unicode 6.0.0.
133  2.  Two minor typos in pcre_internal.h have been fixed.  2.  Two minor typos in pcre_internal.h have been fixed.
135  3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and  3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
136      pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()      pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
137      in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).      in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
139    4.  There were a number of related bugs in the code for matching backrefences
140        caselessly in UTF-8 mode when codes for the characters concerned were
141        different numbers of bytes. For example, U+023A and U+2C65 are an upper
142        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
143        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
144        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
145        2-byte code at the end of the subject (it thought there wasn't enough data
146        left).
148    5.  Comprehensive information about what went wrong is now returned by
149        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
150        as the output vector has at least 2 elements. The offset of the start of
151        the failing character and a reason code are placed in the vector.
153    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
154        now returned is for the first byte of the failing character, instead of the
155        last byte inspected. This is an incompatible change, but I hope it is small
156        enough not to be a problem. It makes the returned offset consistent with
157        pcre_exec() and pcre_dfa_exec().
159    7.  pcretest now gives a text phrase as well as the error number when
160        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
161        failure, the offset and reason code are output.
163    8.  When \R was used with a maximizing quantifier it failed to skip backwards
164        over a \r\n pair if the subsequent match failed. Instead, it just skipped
165        back over a single character (\n). This seems wrong (because it treated the
166        two characters as a single entity when going forwards), conflicts with the
167        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
168        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
169        has been changed.
171    9.  Some internal refactoring has changed the processing so that the handling
172        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
173        time (the PCRE_DOTALL option was changed this way some time ago: version
174        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
175        which was always a bit of a fudge. It also means that there is one less
176        argument for the match() function, which reduces its stack requirements
177        slightly. This change also fixes an incompatibility with Perl: the pattern
178        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
180    10. More internal refactoring has drastically reduced the number of recursive
181        calls to match() for possessively repeated groups such as (abc)++ when
182        using pcre_exec().
184    11. While implementing 10, a number of bugs in the handling of groups were
185        discovered and fixed:
187        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
188        (a|)*(?1) gave a compile-time internal error.
189        ((a|)+)+  did not notice that the outer group could match an empty string.
190        (^a|^)+   was not marked as anchored.
191        (.*a|.*)+ was not marked as matching at start or after a newline.
193    12. Yet more internal refactoring has removed another argument from the match()
194        function. Special calls to this function are now indicated by setting a
195        value in a variable in the "match data" data block.
197    13. Be more explicit in pcre_study() instead of relying on "default" for
198        opcodes that mean there is no starting character; this means that when new
199        ones are added and accidentally left out of pcre_study(), testing should
200        pick them up.
202    14. The -s option of pcretest has been documented for ages as being an old
203        synonym of -m (show memory usage). I have changed it to mean "force study
204        for every regex", that is, assume /S for every regex. This is similar to -i
205        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
206        using it. It makes it easier to run collections of tests with and without
207        study enabled, and thereby test pcre_study() more easily. All the standard
208        tests are now run with and without -s (but some patterns can be marked as
209        "never study" - see 20 below).
211    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
212        restoration of the capturing data to the outer values was not happening
213        correctly.
215    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
216        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
217        pattern had matched an empty string, and so incorrectly returned a no
218        match.
220    17. There was optimizing code for the last branch of non-capturing parentheses,
221        and also for the obeyed branch of a conditional subexpression, which used
222        tail recursion to cut down on stack usage. Unfortunately, now that there is
223        the possibility of (*THEN) occurring in these branches, tail recursion is
224        no longer possible because the return has to be checked for (*THEN). These
225        two optimizations have therefore been removed. [But see 8.20/11 above.]
227    18. If a pattern containing \R was studied, it was assumed that \R always
228        matched two bytes, thus causing the minimum subject length to be
229        incorrectly computed because \R can also match just one byte.
231    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
232        was incorrectly computed.
234    20. If /S is present twice on a test pattern in pcretest input, it now
235        *disables* studying, thereby overriding the use of -s on the command line
236        (see 14 above). This is necessary for one or two tests to keep the output
237        identical in both cases.
239    21. When (*ACCEPT) was used in an assertion that matched an empty string and
240        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
242    22. When an atomic group that contained a capturing parenthesis was
243        successfully matched, but the branch in which it appeared failed, the
244        capturing was not being forgotten if a higher numbered group was later
245        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
246        group 1 to "a", when in fact it should be unset. This applied to multi-
247        branched capturing and non-capturing groups, repeated or not, and also to
248        positive assertions (capturing in negative assertions does not happen
249        in PCRE) and also to nested atomic groups.
251    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
252        subject after a captured substring, to make it easier to tell which of a
253        number of identical substrings has been captured.
255    24. The way atomic groups are processed by pcre_exec() has been changed so that
256        if they are repeated, backtracking one repetition now resets captured
257        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
258        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
259        "aaa". Previously, it would have been "a". As part of this code
260        refactoring, the way recursive calls are handled has also been changed.
262    25. If an assertion condition captured any substrings, they were not passed
263        back unless some other capturing happened later. For example, if
264        (?(?=(a))a) was matched against "a", no capturing was returned.
266    26. When studying a pattern that contained subroutine calls or assertions,
267        the code for finding the minimum length of a possible match was handling
268        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
269        group 1 called group 2 while simultaneously a separate group 2 called group
270        1). A stack overflow occurred in this case. I have fixed this by limiting
271        the recursion depth to 10.
273    27. Updated RunTest.bat in the distribution to the version supplied by Tom
274        Fortmann. This supports explicit test numbers on the command line, and has
275        argument validation and error reporting.
277    28. An instance of \X with an unlimited repeat could fail if at any point the
278        first character it looked at was a mark character.
280    29. Some minor code refactoring concerning Unicode properties and scripts
281        should reduce the stack requirement of match() slightly.
283    30. Added the '=' option to pcretest to check the setting of unused capturing
284        slots at the end of the pattern, which are documented as being -1, but are
285        not included in the return count.
287    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
288        compiled something random. Now it gives a compile-time error (as does
289        Perl).
291    32. A *MARK encountered during the processing of a positive assertion is now
292        recorded and passed back (compatible with Perl).
294    33. If --only-matching or --colour was set on a pcregrep call whose pattern
295        had alternative anchored branches, the search for a second match in a line
296        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
297        matched the line "0102" twice. The same bug affected patterns that started
298        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
299        twice.
301    34. Previously, PCRE did not allow quantification of assertions. However, Perl
302        does, and because of capturing effects, quantifying parenthesized
303        assertions may at times be useful. Quantifiers are now allowed for
304        parenthesized assertions.
306    35. A minor code tidy in pcre_compile() when checking options for \R usage.
308    36. \g was being checked for fancy things in a character class, when it should
309        just be a literal "g".
311    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
312        appearance of a nested POSIX class supersedes an apparent external class.
313        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
314        unescaped square brackets may also appear as part of class names. For
315        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
316        more like Perl. (But see 8.20/1 above.)
318    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
319        was because it thought it was \N{name}, which is not supported).
321    39. Add minix to OS list not supporting the -S option in pcretest.
323    40. PCRE tries to detect cases of infinite recursion at compile time, but it
324        cannot analyze patterns in sufficient detail to catch mutual recursions
325        such as ((?1))((?2)). There is now a runtime test that gives an error if a
326        subgroup is called recursively as a subpattern for a second time at the
327        same position in the subject string. In previous releases this might have
328        been caught by the recursion limit, or it might have run out of stack.
330    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
331        happen only once. PCRE was, however incorrectly giving a compile time error
332        "recursive call could loop indefinitely" because it cannot analyze the
333        pattern in sufficient detail. The compile time test no longer happens when
334        PCRE is compiling a conditional subpattern, but actual runaway loops are
335        now caught at runtime (see 40 above).
337    42. It seems that Perl allows any characters other than a closing parenthesis
338        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
339        has been changed to be the same.
341    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
342        as not to get warnings when autogen.sh is called. Also changed
343        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
345    44. To help people who use pcregrep to scan files containing exceedingly long
346        lines, the following changes have been made:
348        (a) The default value of the buffer size parameter has been increased from
349            8K to 20K. (The actual buffer used is three times this size.)
351        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
352            PCRE is built.
354        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
355            to be set at run time.
357        (d) Numerical values in pcregrep options can be followed by K or M, for
358            example --buffer-size=50K.
360        (e) If a line being scanned overflows pcregrep's buffer, an error is now
361            given and the return code is set to 2.
363    45. Add a pointer to the latest mark to the callout data block.
365    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
366        partial match of an empty string instead of no match. This was specific to
367        the use of ".".
369    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
370        complete match instead of a partial match. This bug was dependent on both
371        the PCRE_UTF8 and PCRE_DOTALL options being set.
373    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
374        starting byte set, because \b was not being ignored.
377  Version 8.12 15-Jan-2011  Version 8.12 15-Jan-2011

Removed from v.594  
changed lines
  Added in v.741

  ViewVC Help
Powered by ViewVC 1.1.5