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

Diff of /code/trunk/ChangeLog

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

revision 595 by ph10, Mon May 2 10:33:29 2011 UTC revision 742 by zherczeg, Sun Nov 6 08:05:33 2011 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 8.13 30-Apr-2011  Version 8.21
5    ------------
6    
7    1.  Updating the JIT compiler.
8    
9    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
10        are added as well.
11    
12    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
13        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
14        calling _pcre_jit_exec. Some extra comments are added.
15    
16    
17    Version 8.20 21-Oct-2011
18    ------------------------
19    
20    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
21        a POSIX class. After further experiments with Perl, which convinced me that
22        Perl has bugs and confusions, a closing square bracket is no longer allowed
23        in a POSIX name. This bug also affected patterns with classes that started
24        with full stops.
25    
26    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
27        captured substring, but while checking the failing first alternative,
28        substring 1 is temporarily captured. If the output vector supplied to
29        pcre_exec() was not big enough for this capture, the yield of the function
30        was still zero ("insufficient space for captured substrings"). This cannot
31        be totally fixed without adding another stack variable, which seems a lot
32        of expense for a edge case. However, I have improved the situation in cases
33        such as /(a)(b)x|abc/ matched against "abc", where the return code
34        indicates that fewer than the maximum number of slots in the ovector have
35        been set.
36    
37    3.  Related to (2) above: when there are more back references in a pattern than
38        slots in the output vector, pcre_exec() uses temporary memory during
39        matching, and copies in the captures as far as possible afterwards. It was
40        using the entire output vector, but this conflicts with the specification
41        that only 2/3 is used for passing back captured substrings. Now it uses
42        only the first 2/3, for compatibility. This is, of course, another edge
43        case.
44    
45    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
46        main code base, and can be used by building with --enable-jit. When this is
47        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
48        runtime --no-jit option is given.
49    
50    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
51        ovector, the return from the function was zero, implying that there were
52        other matches that did not fit. The correct "exactly full" value is now
53        returned.
54    
55    6.  If a subpattern that was called recursively or as a subroutine contained
56        (*PRUNE) or any other control that caused it to give a non-standard return,
57        invalid errors such as "Error -26 (nested recursion at the same subject
58        position)" or even infinite loops could occur.
59    
60    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
61        computing the minimum length on reaching *ACCEPT, and so ended up with the
62        wrong value of 1 rather than 0. Further investigation indicates that
63        computing a minimum subject length in the presence of *ACCEPT is difficult
64        (think back references, subroutine calls), and so I have changed the code
65        so that no minimum is registered for a pattern that contains *ACCEPT.
66    
67    8.  If (*THEN) was present in the first (true) branch of a conditional group,
68        it was not handled as intended. [But see 16 below.]
69    
70    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
71        Sheri Pierce.
72    
73    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
74        the first byte in a match must be "a".
75    
76    11. Change 17 for 8.13 increased the recursion depth for patterns like
77        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
78        pattern contains any instances of (*THEN). If it does not, the old
79        optimizations are restored. It would be nice to do this on a per-group
80        basis, but at the moment that is not feasible.
81    
82    12. In some environments, the output of pcretest -C is CRLF terminated. This
83        broke RunTest's code that checks for the link size. A single white space
84        character after the value is now allowed for.
85    
86    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
87        For "fr", it uses the Windows-specific input and output files.
88    
89    14. If (*THEN) appeared in a group that was called recursively or as a
90        subroutine, it did not work as intended. [But see next item.]
91    
92    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
93        pattern fragments (but not containing any | characters). If A and B are
94        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
95        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
96        D. In other words, Perl considers parentheses that do not contain any |
97        characters to be part of a surrounding alternative, whereas PCRE was
98        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
99        differently. PCRE now behaves in the same way as Perl, except in the case
100        of subroutine/recursion calls such as (?1) which have in any case always
101        been different (but PCRE had them first :-).
102    
103    16. Related to 15 above: Perl does not treat the | in a conditional group as
104        creating alternatives. Such a group is treated in the same way as an
105        ordinary group without any | characters when processing (*THEN). PCRE has
106        been changed to match Perl's behaviour.
107    
108    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
109        RunGrepTest script failed.
110    
111    18. Change 22 for version 13 caused atomic groups to use more stack. This is
112        inevitable for groups that contain captures, but it can lead to a lot of
113        stack use in large patterns. The old behaviour has been restored for atomic
114        groups that do not contain any capturing parentheses.
115    
116    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
117        suppress the check for a minimum subject length at run time. (If it was
118        given to pcre_exec() or pcre_dfa_exec() it did work.)
119    
120    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
121        fail to work when decoding hex characters in data strings in EBCDIC
122        environments.
123    
124    21. It appears that in at least one Mac OS environment, the isxdigit() function
125        is implemented as a macro that evaluates to its argument more than once,
126        contravening the C 90 Standard (I haven't checked a later standard). There
127        was an instance in pcretest which caused it to go wrong when processing
128        \x{...} escapes in subject strings. The has been rewritten to avoid using
129        things like p++ in the argument of isxdigit().
130    
131    
132    Version 8.13 16-Aug-2011
133  ------------------------  ------------------------
134    
135  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.
# Line 21  Version 8.13 30-Apr-2011 Line 149  Version 8.13 30-Apr-2011
149      2-byte code at the end of the subject (it thought there wasn't enough data      2-byte code at the end of the subject (it thought there wasn't enough data
150      left).      left).
151    
152    5.  Comprehensive information about what went wrong is now returned by
153        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
154        as the output vector has at least 2 elements. The offset of the start of
155        the failing character and a reason code are placed in the vector.
156    
157    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
158        now returned is for the first byte of the failing character, instead of the
159        last byte inspected. This is an incompatible change, but I hope it is small
160        enough not to be a problem. It makes the returned offset consistent with
161        pcre_exec() and pcre_dfa_exec().
162    
163    7.  pcretest now gives a text phrase as well as the error number when
164        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
165        failure, the offset and reason code are output.
166    
167    8.  When \R was used with a maximizing quantifier it failed to skip backwards
168        over a \r\n pair if the subsequent match failed. Instead, it just skipped
169        back over a single character (\n). This seems wrong (because it treated the
170        two characters as a single entity when going forwards), conflicts with the
171        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
172        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
173        has been changed.
174    
175    9.  Some internal refactoring has changed the processing so that the handling
176        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
177        time (the PCRE_DOTALL option was changed this way some time ago: version
178        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
179        which was always a bit of a fudge. It also means that there is one less
180        argument for the match() function, which reduces its stack requirements
181        slightly. This change also fixes an incompatibility with Perl: the pattern
182        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
183    
184    10. More internal refactoring has drastically reduced the number of recursive
185        calls to match() for possessively repeated groups such as (abc)++ when
186        using pcre_exec().
187    
188    11. While implementing 10, a number of bugs in the handling of groups were
189        discovered and fixed:
190    
191        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
192        (a|)*(?1) gave a compile-time internal error.
193        ((a|)+)+  did not notice that the outer group could match an empty string.
194        (^a|^)+   was not marked as anchored.
195        (.*a|.*)+ was not marked as matching at start or after a newline.
196    
197    12. Yet more internal refactoring has removed another argument from the match()
198        function. Special calls to this function are now indicated by setting a
199        value in a variable in the "match data" data block.
200    
201    13. Be more explicit in pcre_study() instead of relying on "default" for
202        opcodes that mean there is no starting character; this means that when new
203        ones are added and accidentally left out of pcre_study(), testing should
204        pick them up.
205    
206    14. The -s option of pcretest has been documented for ages as being an old
207        synonym of -m (show memory usage). I have changed it to mean "force study
208        for every regex", that is, assume /S for every regex. This is similar to -i
209        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
210        using it. It makes it easier to run collections of tests with and without
211        study enabled, and thereby test pcre_study() more easily. All the standard
212        tests are now run with and without -s (but some patterns can be marked as
213        "never study" - see 20 below).
214    
215    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
216        restoration of the capturing data to the outer values was not happening
217        correctly.
218    
219    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
220        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
221        pattern had matched an empty string, and so incorrectly returned a no
222        match.
223    
224    17. There was optimizing code for the last branch of non-capturing parentheses,
225        and also for the obeyed branch of a conditional subexpression, which used
226        tail recursion to cut down on stack usage. Unfortunately, now that there is
227        the possibility of (*THEN) occurring in these branches, tail recursion is
228        no longer possible because the return has to be checked for (*THEN). These
229        two optimizations have therefore been removed. [But see 8.20/11 above.]
230    
231    18. If a pattern containing \R was studied, it was assumed that \R always
232        matched two bytes, thus causing the minimum subject length to be
233        incorrectly computed because \R can also match just one byte.
234    
235    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
236        was incorrectly computed.
237    
238    20. If /S is present twice on a test pattern in pcretest input, it now
239        *disables* studying, thereby overriding the use of -s on the command line
240        (see 14 above). This is necessary for one or two tests to keep the output
241        identical in both cases.
242    
243    21. When (*ACCEPT) was used in an assertion that matched an empty string and
244        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
245    
246    22. When an atomic group that contained a capturing parenthesis was
247        successfully matched, but the branch in which it appeared failed, the
248        capturing was not being forgotten if a higher numbered group was later
249        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
250        group 1 to "a", when in fact it should be unset. This applied to multi-
251        branched capturing and non-capturing groups, repeated or not, and also to
252        positive assertions (capturing in negative assertions does not happen
253        in PCRE) and also to nested atomic groups.
254    
255    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
256        subject after a captured substring, to make it easier to tell which of a
257        number of identical substrings has been captured.
258    
259    24. The way atomic groups are processed by pcre_exec() has been changed so that
260        if they are repeated, backtracking one repetition now resets captured
261        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
262        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
263        "aaa". Previously, it would have been "a". As part of this code
264        refactoring, the way recursive calls are handled has also been changed.
265    
266    25. If an assertion condition captured any substrings, they were not passed
267        back unless some other capturing happened later. For example, if
268        (?(?=(a))a) was matched against "a", no capturing was returned.
269    
270    26. When studying a pattern that contained subroutine calls or assertions,
271        the code for finding the minimum length of a possible match was handling
272        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
273        group 1 called group 2 while simultaneously a separate group 2 called group
274        1). A stack overflow occurred in this case. I have fixed this by limiting
275        the recursion depth to 10.
276    
277    27. Updated RunTest.bat in the distribution to the version supplied by Tom
278        Fortmann. This supports explicit test numbers on the command line, and has
279        argument validation and error reporting.
280    
281    28. An instance of \X with an unlimited repeat could fail if at any point the
282        first character it looked at was a mark character.
283    
284    29. Some minor code refactoring concerning Unicode properties and scripts
285        should reduce the stack requirement of match() slightly.
286    
287    30. Added the '=' option to pcretest to check the setting of unused capturing
288        slots at the end of the pattern, which are documented as being -1, but are
289        not included in the return count.
290    
291    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
292        compiled something random. Now it gives a compile-time error (as does
293        Perl).
294    
295    32. A *MARK encountered during the processing of a positive assertion is now
296        recorded and passed back (compatible with Perl).
297    
298    33. If --only-matching or --colour was set on a pcregrep call whose pattern
299        had alternative anchored branches, the search for a second match in a line
300        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
301        matched the line "0102" twice. The same bug affected patterns that started
302        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
303        twice.
304    
305    34. Previously, PCRE did not allow quantification of assertions. However, Perl
306        does, and because of capturing effects, quantifying parenthesized
307        assertions may at times be useful. Quantifiers are now allowed for
308        parenthesized assertions.
309    
310    35. A minor code tidy in pcre_compile() when checking options for \R usage.
311    
312    36. \g was being checked for fancy things in a character class, when it should
313        just be a literal "g".
314    
315    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
316        appearance of a nested POSIX class supersedes an apparent external class.
317        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
318        unescaped square brackets may also appear as part of class names. For
319        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
320        more like Perl. (But see 8.20/1 above.)
321    
322    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
323        was because it thought it was \N{name}, which is not supported).
324    
325    39. Add minix to OS list not supporting the -S option in pcretest.
326    
327    40. PCRE tries to detect cases of infinite recursion at compile time, but it
328        cannot analyze patterns in sufficient detail to catch mutual recursions
329        such as ((?1))((?2)). There is now a runtime test that gives an error if a
330        subgroup is called recursively as a subpattern for a second time at the
331        same position in the subject string. In previous releases this might have
332        been caught by the recursion limit, or it might have run out of stack.
333    
334    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
335        happen only once. PCRE was, however incorrectly giving a compile time error
336        "recursive call could loop indefinitely" because it cannot analyze the
337        pattern in sufficient detail. The compile time test no longer happens when
338        PCRE is compiling a conditional subpattern, but actual runaway loops are
339        now caught at runtime (see 40 above).
340    
341    42. It seems that Perl allows any characters other than a closing parenthesis
342        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
343        has been changed to be the same.
344    
345    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
346        as not to get warnings when autogen.sh is called. Also changed
347        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
348    
349    44. To help people who use pcregrep to scan files containing exceedingly long
350        lines, the following changes have been made:
351    
352        (a) The default value of the buffer size parameter has been increased from
353            8K to 20K. (The actual buffer used is three times this size.)
354    
355        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
356            PCRE is built.
357    
358        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
359            to be set at run time.
360    
361        (d) Numerical values in pcregrep options can be followed by K or M, for
362            example --buffer-size=50K.
363    
364        (e) If a line being scanned overflows pcregrep's buffer, an error is now
365            given and the return code is set to 2.
366    
367    45. Add a pointer to the latest mark to the callout data block.
368    
369    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
370        partial match of an empty string instead of no match. This was specific to
371        the use of ".".
372    
373    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
374        complete match instead of a partial match. This bug was dependent on both
375        the PCRE_UTF8 and PCRE_DOTALL options being set.
376    
377    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
378        starting byte set, because \b was not being ignored.
379    
380    
381  Version 8.12 15-Jan-2011  Version 8.12 15-Jan-2011
382  ------------------------  ------------------------

Legend:
Removed from v.595  
changed lines
  Added in v.742

  ViewVC Help
Powered by ViewVC 1.1.5