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

Diff of /code/trunk/ChangeLog

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

revision 581 by ph10, Fri Dec 10 11:40:00 2010 UTC revision 735 by ph10, Thu Oct 13 15:51:27 2011 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.20 10-Oct-2011
5    ------------------------
6    
7    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
8        a POSIX class. After further experiments with Perl, which convinced me that
9        Perl has bugs and confusions, a closing square bracket is no longer allowed
10        in a POSIX name. This bug also affected patterns with classes that started
11        with full stops.
12    
13    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
14        captured substring, but while checking the failing first alternative,
15        substring 1 is temporarily captured. If the output vector supplied to
16        pcre_exec() was not big enough for this capture, the yield of the function
17        was still zero ("insufficient space for captured substrings"). This cannot
18        be totally fixed without adding another stack variable, which seems a lot
19        of expense for a edge case. However, I have improved the situation in cases
20        such as /(a)(b)x|abc/ matched against "abc", where the return code
21        indicates that fewer than the maximum number of slots in the ovector have
22        been set.
23    
24    3.  Related to (2) above: when there are more back references in a pattern than
25        slots in the output vector, pcre_exec() uses temporary memory during
26        matching, and copies in the captures as far as possible afterwards. It was
27        using the entire output vector, but this conflicts with the specification
28        that only 2/3 is used for passing back captured substrings. Now it uses
29        only the first 2/3, for compatibility. This is, of course, another edge
30        case.
31    
32    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
33        main code base, and can be used by building with --enable-jit. When this is
34        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
35        runtime --no-jit option is given.
36    
37    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
38        ovector, the return from the function was zero, implying that there were
39        other matches that did not fit. The correct "exactly full" value is now
40        returned.
41    
42    6.  If a subpattern that was called recursively or as a subroutine contained
43        (*PRUNE) or any other control that caused it to give a non-standard return,
44        invalid errors such as "Error -26 (nested recursion at the same subject
45        position)" or even infinite loops could occur.
46    
47    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
48        computing the minimum length on reaching *ACCEPT, and so ended up with the
49        wrong value of 1 rather than 0. Further investigation indicates that
50        computing a minimum subject length in the presence of *ACCEPT is difficult
51        (think back references, subroutine calls), and so I have changed the code
52        so that no minimum is registered for a pattern that contains *ACCEPT.
53    
54    8.  If (*THEN) was present in the first (true) branch of a conditional group,
55        it was not handled as intended. [But see 16 below.]
56    
57    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
58        Sheri Pierce.
59    
60    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
61        the first byte in a match must be "a".
62    
63    11. Change 17 for 8.13 increased the recursion depth for patterns like
64        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
65        pattern contains any instances of (*THEN). If it does not, the old
66        optimizations are restored. It would be nice to do this on a per-group
67        basis, but at the moment that is not feasible.
68    
69    12. In some environments, the output of pcretest -C is CRLF terminated. This
70        broke RunTest's code that checks for the link size. A single white space
71        after the value is now allowed for.
72    
73    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
74        For "fr", it uses the Windows-specific input and output files.
75    
76    14. If (*THEN) appeared in a group that was called recursively or as a
77        subroutine, it did not work as intended. [But see next item.]
78    
79    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
80        pattern fragments (but not containing any | characters). If A and B are
81        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
82        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
83        D. In other words, Perl considers parentheses that do not contain any |
84        characters to be part of a surrounding alternative, whereas PCRE was
85        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
86        differently. PCRE now behaves in the same way as Perl, except in the case
87        of subroutine/recursion calls such as (?1) which have in any case always
88        been different (but PCRE had them first :-).
89    
90    16. Related to 15 above: Perl does not treat the | in a conditional group as
91        creating alternatives. Such a group is treated in the same way as an
92        ordinary group without any | characters when processing (*THEN). PCRE has
93        been changed to match Perl's behaviour.
94    
95    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
96        RunGrepTest script failed.
97    
98    18. Change 22 for version 13 caused atomic groups to use more stack. This is
99        inevitable for groups that contain captures, but it can lead to a lot of
100        stack use in large patterns. The old behaviour has been restored for atomic
101        groups that do not contain any capturing parentheses.
102    
103    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
104        suppress the check for a minimum subject length at run time. (If it was
105        given to pcre_exec() or pcre_dfa_exec() it did work.)
106    
107    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
108        fail to work when decoding hex characters in data strings in EBCDIC
109        environments.
110    
111    21. It appears that in at least one Mac OS environment, the isxdigit() function
112        is implemented as a macro that evaluates to its argument more than once,
113        contravening the C 90 Standard (I haven't checked a later standard). There
114        was an instance in pcretest which caused it to go wrong when processing
115        \x{...} escapes in subject strings. The has been rewritten to avoid using
116        things like p++ in the argument of isxdigit().
117    
118    
119    Version 8.13 16-Aug-2011
120    ------------------------
121    
122    1.  The Unicode data tables have been updated to Unicode 6.0.0.
123    
124    2.  Two minor typos in pcre_internal.h have been fixed.
125    
126    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
127        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
128        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
129    
130    4.  There were a number of related bugs in the code for matching backrefences
131        caselessly in UTF-8 mode when codes for the characters concerned were
132        different numbers of bytes. For example, U+023A and U+2C65 are an upper
133        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
134        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
135        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
136        2-byte code at the end of the subject (it thought there wasn't enough data
137        left).
138    
139    5.  Comprehensive information about what went wrong is now returned by
140        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
141        as the output vector has at least 2 elements. The offset of the start of
142        the failing character and a reason code are placed in the vector.
143    
144    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
145        now returned is for the first byte of the failing character, instead of the
146        last byte inspected. This is an incompatible change, but I hope it is small
147        enough not to be a problem. It makes the returned offset consistent with
148        pcre_exec() and pcre_dfa_exec().
149    
150    7.  pcretest now gives a text phrase as well as the error number when
151        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
152        failure, the offset and reason code are output.
153    
154    8.  When \R was used with a maximizing quantifier it failed to skip backwards
155        over a \r\n pair if the subsequent match failed. Instead, it just skipped
156        back over a single character (\n). This seems wrong (because it treated the
157        two characters as a single entity when going forwards), conflicts with the
158        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
159        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
160        has been changed.
161    
162    9.  Some internal refactoring has changed the processing so that the handling
163        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
164        time (the PCRE_DOTALL option was changed this way some time ago: version
165        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
166        which was always a bit of a fudge. It also means that there is one less
167        argument for the match() function, which reduces its stack requirements
168        slightly. This change also fixes an incompatibility with Perl: the pattern
169        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
170    
171    10. More internal refactoring has drastically reduced the number of recursive
172        calls to match() for possessively repeated groups such as (abc)++ when
173        using pcre_exec().
174    
175    11. While implementing 10, a number of bugs in the handling of groups were
176        discovered and fixed:
177    
178        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
179        (a|)*(?1) gave a compile-time internal error.
180        ((a|)+)+  did not notice that the outer group could match an empty string.
181        (^a|^)+   was not marked as anchored.
182        (.*a|.*)+ was not marked as matching at start or after a newline.
183    
184    12. Yet more internal refactoring has removed another argument from the match()
185        function. Special calls to this function are now indicated by setting a
186        value in a variable in the "match data" data block.
187    
188    13. Be more explicit in pcre_study() instead of relying on "default" for
189        opcodes that mean there is no starting character; this means that when new
190        ones are added and accidentally left out of pcre_study(), testing should
191        pick them up.
192    
193    14. The -s option of pcretest has been documented for ages as being an old
194        synonym of -m (show memory usage). I have changed it to mean "force study
195        for every regex", that is, assume /S for every regex. This is similar to -i
196        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
197        using it. It makes it easier to run collections of tests with and without
198        study enabled, and thereby test pcre_study() more easily. All the standard
199        tests are now run with and without -s (but some patterns can be marked as
200        "never study" - see 20 below).
201    
202    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
203        restoration of the capturing data to the outer values was not happening
204        correctly.
205    
206    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
207        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
208        pattern had matched an empty string, and so incorrectly returned a no
209        match.
210    
211    17. There was optimizing code for the last branch of non-capturing parentheses,
212        and also for the obeyed branch of a conditional subexpression, which used
213        tail recursion to cut down on stack usage. Unfortunately, now that there is
214        the possibility of (*THEN) occurring in these branches, tail recursion is
215        no longer possible because the return has to be checked for (*THEN). These
216        two optimizations have therefore been removed. [But see 8.20/11 above.]
217    
218    18. If a pattern containing \R was studied, it was assumed that \R always
219        matched two bytes, thus causing the minimum subject length to be
220        incorrectly computed because \R can also match just one byte.
221    
222    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
223        was incorrectly computed.
224    
225    20. If /S is present twice on a test pattern in pcretest input, it now
226        *disables* studying, thereby overriding the use of -s on the command line
227        (see 14 above). This is necessary for one or two tests to keep the output
228        identical in both cases.
229    
230    21. When (*ACCEPT) was used in an assertion that matched an empty string and
231        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
232    
233    22. When an atomic group that contained a capturing parenthesis was
234        successfully matched, but the branch in which it appeared failed, the
235        capturing was not being forgotten if a higher numbered group was later
236        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
237        group 1 to "a", when in fact it should be unset. This applied to multi-
238        branched capturing and non-capturing groups, repeated or not, and also to
239        positive assertions (capturing in negative assertions does not happen
240        in PCRE) and also to nested atomic groups.
241    
242    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
243        subject after a captured substring, to make it easier to tell which of a
244        number of identical substrings has been captured.
245    
246    24. The way atomic groups are processed by pcre_exec() has been changed so that
247        if they are repeated, backtracking one repetition now resets captured
248        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
249        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
250        "aaa". Previously, it would have been "a". As part of this code
251        refactoring, the way recursive calls are handled has also been changed.
252    
253    25. If an assertion condition captured any substrings, they were not passed
254        back unless some other capturing happened later. For example, if
255        (?(?=(a))a) was matched against "a", no capturing was returned.
256    
257    26. When studying a pattern that contained subroutine calls or assertions,
258        the code for finding the minimum length of a possible match was handling
259        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
260        group 1 called group 2 while simultaneously a separate group 2 called group
261        1). A stack overflow occurred in this case. I have fixed this by limiting
262        the recursion depth to 10.
263    
264    27. Updated RunTest.bat in the distribution to the version supplied by Tom
265        Fortmann. This supports explicit test numbers on the command line, and has
266        argument validation and error reporting.
267    
268    28. An instance of \X with an unlimited repeat could fail if at any point the
269        first character it looked at was a mark character.
270    
271    29. Some minor code refactoring concerning Unicode properties and scripts
272        should reduce the stack requirement of match() slightly.
273    
274    30. Added the '=' option to pcretest to check the setting of unused capturing
275        slots at the end of the pattern, which are documented as being -1, but are
276        not included in the return count.
277    
278    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
279        compiled something random. Now it gives a compile-time error (as does
280        Perl).
281    
282    32. A *MARK encountered during the processing of a positive assertion is now
283        recorded and passed back (compatible with Perl).
284    
285    33. If --only-matching or --colour was set on a pcregrep call whose pattern
286        had alternative anchored branches, the search for a second match in a line
287        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
288        matched the line "0102" twice. The same bug affected patterns that started
289        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
290        twice.
291    
292    34. Previously, PCRE did not allow quantification of assertions. However, Perl
293        does, and because of capturing effects, quantifying parenthesized
294        assertions may at times be useful. Quantifiers are now allowed for
295        parenthesized assertions.
296    
297    35. A minor code tidy in pcre_compile() when checking options for \R usage.
298    
299    36. \g was being checked for fancy things in a character class, when it should
300        just be a literal "g".
301    
302    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
303        appearance of a nested POSIX class supersedes an apparent external class.
304        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
305        unescaped square brackets may also appear as part of class names. For
306        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
307        more like Perl. (But see 8.20/1 above.)
308    
309    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
310        was because it thought it was \N{name}, which is not supported).
311    
312    39. Add minix to OS list not supporting the -S option in pcretest.
313    
314    40. PCRE tries to detect cases of infinite recursion at compile time, but it
315        cannot analyze patterns in sufficient detail to catch mutual recursions
316        such as ((?1))((?2)). There is now a runtime test that gives an error if a
317        subgroup is called recursively as a subpattern for a second time at the
318        same position in the subject string. In previous releases this might have
319        been caught by the recursion limit, or it might have run out of stack.
320    
321    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
322        happen only once. PCRE was, however incorrectly giving a compile time error
323        "recursive call could loop indefinitely" because it cannot analyze the
324        pattern in sufficient detail. The compile time test no longer happens when
325        PCRE is compiling a conditional subpattern, but actual runaway loops are
326        now caught at runtime (see 40 above).
327    
328    42. It seems that Perl allows any characters other than a closing parenthesis
329        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
330        has been changed to be the same.
331    
332    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
333        as not to get warnings when autogen.sh is called. Also changed
334        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
335    
336    44. To help people who use pcregrep to scan files containing exceedingly long
337        lines, the following changes have been made:
338    
339        (a) The default value of the buffer size parameter has been increased from
340            8K to 20K. (The actual buffer used is three times this size.)
341    
342        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
343            PCRE is built.
344    
345        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
346            to be set at run time.
347    
348        (d) Numerical values in pcregrep options can be followed by K or M, for
349            example --buffer-size=50K.
350    
351        (e) If a line being scanned overflows pcregrep's buffer, an error is now
352            given and the return code is set to 2.
353    
354    45. Add a pointer to the latest mark to the callout data block.
355    
356    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
357        partial match of an empty string instead of no match. This was specific to
358        the use of ".".
359    
360    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
361        complete match instead of a partial match. This bug was dependent on both
362        the PCRE_UTF8 and PCRE_DOTALL options being set.
363    
364    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
365        starting byte set, because \b was not being ignored.
366    
367    
368    Version 8.12 15-Jan-2011
369    ------------------------
370    
371    1.  Fixed some typos in the markup of the man pages, and wrote a script that
372        checks for such things as part of the documentation building process.
373    
374    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
375        --match-limit and --recursion-limit options (added for 8.11). In
376        particular, this made one of the standard tests fail. (The integer value
377        went into the wrong half of a long int.)
378    
379    3.  If the --colour option was given to pcregrep with -v (invert match), it
380        did strange things, either producing crazy output, or crashing. It should,
381        of course, ignore a request for colour when reporting lines that do not
382        match.
383    
384    4.  Another pcregrep bug caused similar problems if --colour was specified with
385        -M (multiline) and the pattern match finished with a line ending.
386    
387    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
388        matched in multiline mode, the following line was shown as part of the
389        match. This seems wrong, so I have changed it.
390    
391    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
392        the check for further matches in the same line (so they could be coloured)
393        to overrun the end of the current line. If another match was found, it was
394        incorrectly shown (and then shown again when found in the next line).
395    
396    7.  If pcregrep was compiled under Windows, there was a reference to the
397        function pcregrep_exit() before it was defined. I am assuming this was
398        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
399        reported by a user. I've moved the definition above the reference.
400    
401    
402  Version 8.11 10-Dec-2010  Version 8.11 10-Dec-2010
403  ------------------------  ------------------------
404    

Legend:
Removed from v.581  
changed lines
  Added in v.735

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12