ViewVC logotype

Diff of /code/trunk/ChangeLog

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

revision 551 by ph10, Sun Oct 10 17:33:07 2010 UTC revision 700 by ph10, Tue Sep 20 10:51:26 2011 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
4  Version 8.11 10-Oct-2010  Version 8.20 12-Sep-2011
5    ------------------------
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.
13    2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no captured
14       substring, but while checking the failing first alternative, substring 1 is
15       temporarily captured. If the output vector supplied to pcre_exec() was not
16       big enough for this capture, the yield of the function was still zero
17       ("insufficient space for captured substrings"). This cannot be totally fixed
18       without adding another stack variable, which seems a lot of expense for a
19       edge case. However, I have improved the situation in cases such as
20       /(a)(b)x|abc/ matched against "abc", where the return code indicates that
21       fewer than the maximum number of slots in the ovector have been set.
23    3. Related to (2) above: when there are more back references in a pattern than
24       slots in the output vector, pcre_exec() uses temporary memory during
25       matching, and copies in the captures as far as possible afterwards. It was
26       using the entire output vector, but this conflicts with the specification
27       that only 2/3 is used for passing back captured substrings. Now it uses only
28       the first 2/3, for compatibility. This is, of course, another edge case.
30    4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
31       main code base, and can be used by building with --enable-jit. When this is
32       done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
33       runtime --no-jit option is given.
35    5. When the number of matches in a pcre_dfa_exec() run exactly filled the
36       ovector, the return from the function was zero, implying that there were
37       other matches that did not fit. The correct "exactly full" value is now
38       returned.
40    6. If a subpattern that was called recursively or as a subroutine contained
41       (*PRUNE) or any other control that caused it to give a non-standard return,
42       invalid errors such as "Error -26 (nested recursion at the same subject
43       position)" or even infinite loops could occur.
45    7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
46       computing the minimum length on reaching *ACCEPT, and so ended up with the
47       wrong value of 1 rather than 0. Further investigation indicates that
48       computing a minimum subject length in the presence of *ACCEPT is difficult
49       (think back references, subroutine calls), and so I have changed the code so
50       that no minimum is registered for a pattern that contains *ACCEPT.
52    8. If (*THEN) was present in the first (true) branch of a conditional group,
53       it was not handled as intended.
55    9. Replaced RunTest.bat with the much improved version provided by Sheri
56       Pierce.
59    Version 8.13 16-Aug-2011
60    ------------------------
62    1.  The Unicode data tables have been updated to Unicode 6.0.0.
64    2.  Two minor typos in pcre_internal.h have been fixed.
66    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
67        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
68        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
70    4.  There were a number of related bugs in the code for matching backrefences
71        caselessly in UTF-8 mode when codes for the characters concerned were
72        different numbers of bytes. For example, U+023A and U+2C65 are an upper
73        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
74        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
75        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
76        2-byte code at the end of the subject (it thought there wasn't enough data
77        left).
79    5.  Comprehensive information about what went wrong is now returned by
80        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
81        as the output vector has at least 2 elements. The offset of the start of
82        the failing character and a reason code are placed in the vector.
84    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
85        now returned is for the first byte of the failing character, instead of the
86        last byte inspected. This is an incompatible change, but I hope it is small
87        enough not to be a problem. It makes the returned offset consistent with
88        pcre_exec() and pcre_dfa_exec().
90    7.  pcretest now gives a text phrase as well as the error number when
91        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
92        failure, the offset and reason code are output.
94    8.  When \R was used with a maximizing quantifier it failed to skip backwards
95        over a \r\n pair if the subsequent match failed. Instead, it just skipped
96        back over a single character (\n). This seems wrong (because it treated the
97        two characters as a single entity when going forwards), conflicts with the
98        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
99        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
100        has been changed.
102    9.  Some internal refactoring has changed the processing so that the handling
103        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
104        time (the PCRE_DOTALL option was changed this way some time ago: version
105        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
106        which was always a bit of a fudge. It also means that there is one less
107        argument for the match() function, which reduces its stack requirements
108        slightly. This change also fixes an incompatibility with Perl: the pattern
109        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
111    10. More internal refactoring has drastically reduced the number of recursive
112        calls to match() for possessively repeated groups such as (abc)++ when
113        using pcre_exec().
115    11. While implementing 10, a number of bugs in the handling of groups were
116        discovered and fixed:
118        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
119        (a|)*(?1) gave a compile-time internal error.
120        ((a|)+)+  did not notice that the outer group could match an empty string.
121        (^a|^)+   was not marked as anchored.
122        (.*a|.*)+ was not marked as matching at start or after a newline.
124    12. Yet more internal refactoring has removed another argument from the match()
125        function. Special calls to this function are now indicated by setting a
126        value in a variable in the "match data" data block.
128    13. Be more explicit in pcre_study() instead of relying on "default" for
129        opcodes that mean there is no starting character; this means that when new
130        ones are added and accidentally left out of pcre_study(), testing should
131        pick them up.
133    14. The -s option of pcretest has been documented for ages as being an old
134        synonym of -m (show memory usage). I have changed it to mean "force study
135        for every regex", that is, assume /S for every regex. This is similar to -i
136        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
137        using it. It makes it easier to run collections of tests with and without
138        study enabled, and thereby test pcre_study() more easily. All the standard
139        tests are now run with and without -s (but some patterns can be marked as
140        "never study" - see 20 below).
142    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
143        restoration of the capturing data to the outer values was not happening
144        correctly.
146    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
147        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
148        pattern had matched an empty string, and so incorrectly returned a no
149        match.
151    17. There was optimizing code for the last branch of non-capturing parentheses,
152        and also for the obeyed branch of a conditional subexpression, which used
153        tail recursion to cut down on stack usage. Unfortunately, now that there is
154        the possibility of (*THEN) occurring in these branches, tail recursion is
155        no longer possible because the return has to be checked for (*THEN). These
156        two optimizations have therefore been removed.
158    18. If a pattern containing \R was studied, it was assumed that \R always
159        matched two bytes, thus causing the minimum subject length to be
160        incorrectly computed because \R can also match just one byte.
162    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
163        was incorrectly computed.
165    20. If /S is present twice on a test pattern in pcretest input, it now
166        *disables* studying, thereby overriding the use of -s on the command line
167        (see 14 above). This is necessary for one or two tests to keep the output
168        identical in both cases.
170    21. When (*ACCEPT) was used in an assertion that matched an empty string and
171        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
173    22. When an atomic group that contained a capturing parenthesis was
174        successfully matched, but the branch in which it appeared failed, the
175        capturing was not being forgotten if a higher numbered group was later
176        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
177        group 1 to "a", when in fact it should be unset. This applied to multi-
178        branched capturing and non-capturing groups, repeated or not, and also to
179        positive assertions (capturing in negative assertions does not happen
180        in PCRE) and also to nested atomic groups.
182    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
183        subject after a captured substring, to make it easier to tell which of a
184        number of identical substrings has been captured.
186    24. The way atomic groups are processed by pcre_exec() has been changed so that
187        if they are repeated, backtracking one repetition now resets captured
188        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
189        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
190        "aaa". Previously, it would have been "a". As part of this code
191        refactoring, the way recursive calls are handled has also been changed.
193    25. If an assertion condition captured any substrings, they were not passed
194        back unless some other capturing happened later. For example, if
195        (?(?=(a))a) was matched against "a", no capturing was returned.
197    26. When studying a pattern that contained subroutine calls or assertions,
198        the code for finding the minimum length of a possible match was handling
199        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
200        group 1 called group 2 while simultaneously a separate group 2 called group
201        1). A stack overflow occurred in this case. I have fixed this by limiting
202        the recursion depth to 10.
204    27. Updated RunTest.bat in the distribution to the version supplied by Tom
205        Fortmann. This supports explicit test numbers on the command line, and has
206        argument validation and error reporting.
208    28. An instance of \X with an unlimited repeat could fail if at any point the
209        first character it looked at was a mark character.
211    29. Some minor code refactoring concerning Unicode properties and scripts
212        should reduce the stack requirement of match() slightly.
214    30. Added the '=' option to pcretest to check the setting of unused capturing
215        slots at the end of the pattern, which are documented as being -1, but are
216        not included in the return count.
218    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
219        compiled something random. Now it gives a compile-time error (as does
220        Perl).
222    32. A *MARK encountered during the processing of a positive assertion is now
223        recorded and passed back (compatible with Perl).
225    33. If --only-matching or --colour was set on a pcregrep call whose pattern
226        had alternative anchored branches, the search for a second match in a line
227        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
228        matched the line "0102" twice. The same bug affected patterns that started
229        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
230        twice.
232    34. Previously, PCRE did not allow quantification of assertions. However, Perl
233        does, and because of capturing effects, quantifying parenthesized
234        assertions may at times be useful. Quantifiers are now allowed for
235        parenthesized assertions.
237    35. A minor code tidy in pcre_compile() when checking options for \R usage.
239    36. \g was being checked for fancy things in a character class, when it should
240        just be a literal "g".
242    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
243        appearance of a nested POSIX class supersedes an apparent external class.
244        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
245        unescaped square brackets may also appear as part of class names. For
246        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
247        more like Perl. (But see 8.20/1 above.)
249    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
250        was because it thought it was \N{name}, which is not supported).
252    39. Add minix to OS list not supporting the -S option in pcretest.
254    40. PCRE tries to detect cases of infinite recursion at compile time, but it
255        cannot analyze patterns in sufficient detail to catch mutual recursions
256        such as ((?1))((?2)). There is now a runtime test that gives an error if a
257        subgroup is called recursively as a subpattern for a second time at the
258        same position in the subject string. In previous releases this might have
259        been caught by the recursion limit, or it might have run out of stack.
261    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
262        happen only once. PCRE was, however incorrectly giving a compile time error
263        "recursive call could loop indefinitely" because it cannot analyze the
264        pattern in sufficient detail. The compile time test no longer happens when
265        PCRE is compiling a conditional subpattern, but actual runaway loops are
266        now caught at runtime (see 40 above).
268    42. It seems that Perl allows any characters other than a closing parenthesis
269        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
270        has been changed to be the same.
272    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
273        as not to get warnings when autogen.sh is called. Also changed
274        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
276    44. To help people who use pcregrep to scan files containing exceedingly long
277        lines, the following changes have been made:
279        (a) The default value of the buffer size parameter has been increased from
280            8K to 20K. (The actual buffer used is three times this size.)
282        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
283            PCRE is built.
285        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
286            to be set at run time.
288        (d) Numerical values in pcregrep options can be followed by K or M, for
289            example --buffer-size=50K.
291        (e) If a line being scanned overflows pcregrep's buffer, an error is now
292            given and the return code is set to 2.
294    45. Add a pointer to the latest mark to the callout data block.
296    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
297        partial match of an empty string instead of no match. This was specific to
298        the use of ".".
300    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
301        complete match instead of a partial match. This bug was dependent on both
302        the PCRE_UTF8 and PCRE_DOTALL options being set.
304    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
305        starting byte set, because \b was not being ignored.
308    Version 8.12 15-Jan-2011
309    ------------------------
311    1.  Fixed some typos in the markup of the man pages, and wrote a script that
312        checks for such things as part of the documentation building process.
314    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
315        --match-limit and --recursion-limit options (added for 8.11). In
316        particular, this made one of the standard tests fail. (The integer value
317        went into the wrong half of a long int.)
319    3.  If the --colour option was given to pcregrep with -v (invert match), it
320        did strange things, either producing crazy output, or crashing. It should,
321        of course, ignore a request for colour when reporting lines that do not
322        match.
324    4.  Another pcregrep bug caused similar problems if --colour was specified with
325        -M (multiline) and the pattern match finished with a line ending.
327    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
328        matched in multiline mode, the following line was shown as part of the
329        match. This seems wrong, so I have changed it.
331    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
332        the check for further matches in the same line (so they could be coloured)
333        to overrun the end of the current line. If another match was found, it was
334        incorrectly shown (and then shown again when found in the next line).
336    7.  If pcregrep was compiled under Windows, there was a reference to the
337        function pcregrep_exit() before it was defined. I am assuming this was
338        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
339        reported by a user. I've moved the definition above the reference.
342    Version 8.11 10-Dec-2010
343  ------------------------  ------------------------
345  1.  (*THEN) was not working properly if there were untried alternatives prior  1.  (*THEN) was not working properly if there were untried alternatives prior
346      to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it      to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
347      backtracked to try for "b" instead of moving to the next alternative branch      backtracked to try for "b" instead of moving to the next alternative branch
348      at the same level (in this case, to look for "c"). The Perl documentation      at the same level (in this case, to look for "c"). The Perl documentation
349      is clear that when (*THEN) is backtracked onto, it goes to the "next      is clear that when (*THEN) is backtracked onto, it goes to the "next
350      alternative in the innermost enclosing group".      alternative in the innermost enclosing group".
352  2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern  2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
353      such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should      such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
354      result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and      result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
355      (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides      (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
356      (*THEN).      (*THEN).
358    3.  If \s appeared in a character class, it removed the VT character from
359        the class, even if it had been included by some previous item, for example
360        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
361        of \s, but is part of the POSIX "space" class.)
363    4.  A partial match never returns an empty string (because you can always
364        match an empty string at the end of the subject); however the checking for
365        an empty string was starting at the "start of match" point. This has been
366        changed to the "earliest inspected character" point, because the returned
367        data for a partial match starts at this character. This means that, for
368        example, /(?<=abc)def/ gives a partial match for the subject "abc"
369        (previously it gave "no match").
371    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
372        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
373        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
374        has an implication that the given string is incomplete (because a partial
375        match is preferred over a full match). For this reason, these items now
376        give a partial match in this situation. [Aside: previously, the one case
377        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
378        match rather than a full match, which was wrong by the old rules, but is
379        now correct.]
381    6.  There was a bug in the handling of #-introduced comments, recognized when
382        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
383        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
384        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
385        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
386        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
387        places in pcre_compile().
389    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
390        comments when looking ahead for named forward references to subpatterns,
391        the only newline sequence it recognized was NL. It now handles newlines
392        according to the set newline convention.
394    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
395        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
396        cater for a lack of strerror(). These oversights have been fixed.
398    9.  Added --match-limit and --recursion-limit to pcregrep.
400    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
402    11. When the -o option was used, pcregrep was setting a return code of 1, even
403        when matches were found, and --line-buffered was not being honoured.
405    12. Added an optional parentheses number to the -o and --only-matching options
406        of pcregrep.
408    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
409        can match an empty string. The code to do it in pcretest and pcredemo
410        needed fixing:
412        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
413            only one byte after an empty string match just before CRLF (this case
414            just got forgotten; "any" and "anycrlf" were OK).
416        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
417            mode when an empty string match preceded an ASCII character followed by
418            a non-ASCII character. (The code for advancing by one character rather
419            than one byte was nonsense.)
421        (c) The pcredemo.c sample program did not have any code at all to handle
422            the cases when CRLF is a valid newline sequence.
424    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
425        as a starting offset was within the subject string. There is now a new
426        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
427        negative or greater than the length of the string. In order to test this,
428        pcretest is extended to allow the setting of negative starting offsets.
430    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
431        starting offset points to the beginning of a UTF-8 character was
432        unnecessarily clumsy. I tidied it up.
434    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
435        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
437    17. Nobody had reported that the --include_dir option, which was added in
438        release 7.7 should have been called --include-dir (hyphen, not underscore)
439        for compatibility with GNU grep. I have changed it to --include-dir, but
440        left --include_dir as an undocumented synonym, and the same for
441        --exclude-dir, though that is not available in GNU grep, at least as of
442        release 2.5.4.
444    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
445        characters from a string of bytes) have been redefined so as not to use
446        loops, in order to improve performance in some environments. At the same
447        time, I abstracted some of the common code into auxiliary macros to save
448        repetition (this should not affect the compiled code).
450    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
451        compile-time error is now given if \c is not followed by an ASCII
452        character, that is, a byte less than 128. (In EBCDIC mode, the code is
453        different, and any byte value is allowed.)
455    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
456        START_OPTIMIZE option, which is now allowed at compile time - but just
457        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
458        to pcregrep and other applications that have no direct access to PCRE
459        options. The new /Y option in pcretest sets this option when calling
460        pcre_compile().
462    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
463        back references. Groups containing recursive back references were forced to
464        be atomic by that change, but in the case of named groups, the amount of
465        memory required was incorrectly computed, leading to "Failed: internal
466        error: code overflow". This has been fixed.
468    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
469        pcretest.c, to avoid build problems in some Borland environments.
472  Version 8.10 25-Jun-2010  Version 8.10 25-Jun-2010

Removed from v.551  
changed lines
  Added in v.700

  ViewVC Help
Powered by ViewVC 1.1.5