/[pcre]/code/branches/pcre16/ChangeLog

Diff of /code/branches/pcre16/ChangeLog

Parent Directory | Revision Log | View Patch Patch

-revision 619 by ph10,
Sun Jul 17 13:23:14 2011 UTC
+revision 752 by zherczeg,
Sat Nov 19 15:28:29 2011 UTC
 Line 1
  ChangeLog for PCRE
  ------------------
- Version 8.13 30-Apr-2011
+ Version 8.21
+ ------------
+.  Updating the JIT compiler.
+.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
+     are added as well.
+.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
+     PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
+     calling _pcre_jit_exec. Some extra comments are added.
+.  Mark settings inside atomic groups that do not contain any capturing
+     parentheses, for example, (?>a(*:m)), were not being passed out. This bug
+     was introduced by change 18 for 8.20.
+.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
+     ECMA-262 standard.
+.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
+     erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
+     This bug was probably introduced by change 9 of 8.13.
+.  While fixing 6 above, I noticed that a number of other items were being
+     incorrectly rejected as "not fixed length". This arose partly because newer
+     opcodes had not been added to the fixed-length checking code. I have (a)
+     corrected the bug and added tests for these items, and (b) arranged for an
+     error to occur if an unknown opcode is encountered while checking for fixed
+     length instead of just assuming "not fixed length". The items that were
+     rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
+     (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
+     repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
+.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
+     being incorrectly compiled and would have given unpredicatble results.
+.  A possessively repeated subpattern with minimum repeat count greater than
+     one behaved incorrectly. For example, (A){2,}+ behaved as if it was
+     (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
+     the first (A) could occur when it should not.
+. Add a cast and remove a redundant test from the code.
+. JIT should use pcre_malloc/pcre_free for allocation.
+ Version 8.20 21-Oct-2011
+ ------------------------
+.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
+     a POSIX class. After further experiments with Perl, which convinced me that
+     Perl has bugs and confusions, a closing square bracket is no longer allowed
+     in a POSIX name. This bug also affected patterns with classes that started
+     with full stops.
+.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
+     captured substring, but while checking the failing first alternative,
+     substring 1 is temporarily captured. If the output vector supplied to
+     pcre_exec() was not big enough for this capture, the yield of the function
+     was still zero ("insufficient space for captured substrings"). This cannot
+     be totally fixed without adding another stack variable, which seems a lot
+     of expense for a edge case. However, I have improved the situation in cases
+     such as /(a)(b)x|abc/ matched against "abc", where the return code
+     indicates that fewer than the maximum number of slots in the ovector have
+     been set.
+.  Related to (2) above: when there are more back references in a pattern than
+     slots in the output vector, pcre_exec() uses temporary memory during
+     matching, and copies in the captures as far as possible afterwards. It was
+     using the entire output vector, but this conflicts with the specification
+     that only 2/3 is used for passing back captured substrings. Now it uses
+     only the first 2/3, for compatibility. This is, of course, another edge
+     case.
+.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
+     main code base, and can be used by building with --enable-jit. When this is
+     done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
+     runtime --no-jit option is given.
+.  When the number of matches in a pcre_dfa_exec() run exactly filled the
+     ovector, the return from the function was zero, implying that there were
+     other matches that did not fit. The correct "exactly full" value is now
+     returned.
+.  If a subpattern that was called recursively or as a subroutine contained
+     (*PRUNE) or any other control that caused it to give a non-standard return,
+     invalid errors such as "Error -26 (nested recursion at the same subject
+     position)" or even infinite loops could occur.
+.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
+     computing the minimum length on reaching *ACCEPT, and so ended up with the
+     wrong value of 1 rather than 0. Further investigation indicates that
+     computing a minimum subject length in the presence of *ACCEPT is difficult
+     (think back references, subroutine calls), and so I have changed the code
+     so that no minimum is registered for a pattern that contains *ACCEPT.
+.  If (*THEN) was present in the first (true) branch of a conditional group,
+     it was not handled as intended. [But see 16 below.]
+.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
+     Sheri Pierce.
+. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
+     the first byte in a match must be "a".
+. Change 17 for 8.13 increased the recursion depth for patterns like
+     /a(?:.)*?a/ drastically. I've improved things by remembering whether a
+     pattern contains any instances of (*THEN). If it does not, the old
+     optimizations are restored. It would be nice to do this on a per-group
+     basis, but at the moment that is not feasible.
+. In some environments, the output of pcretest -C is CRLF terminated. This
+     broke RunTest's code that checks for the link size. A single white space
+     character after the value is now allowed for.
+. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
+     For "fr", it uses the Windows-specific input and output files.
+. If (*THEN) appeared in a group that was called recursively or as a
+     subroutine, it did not work as intended. [But see next item.]
+. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
+     pattern fragments (but not containing any | characters). If A and B are
+     matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
+     was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
+     D. In other words, Perl considers parentheses that do not contain any |
+     characters to be part of a surrounding alternative, whereas PCRE was
+     treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
+     differently. PCRE now behaves in the same way as Perl, except in the case
+     of subroutine/recursion calls such as (?1) which have in any case always
+     been different (but PCRE had them first :-).
+. Related to 15 above: Perl does not treat the | in a conditional group as
+     creating alternatives. Such a group is treated in the same way as an
+     ordinary group without any | characters when processing (*THEN). PCRE has
+     been changed to match Perl's behaviour.
+. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
+     RunGrepTest script failed.
+. Change 22 for version 13 caused atomic groups to use more stack. This is
+     inevitable for groups that contain captures, but it can lead to a lot of
+     stack use in large patterns. The old behaviour has been restored for atomic
+     groups that do not contain any capturing parentheses.
+. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
+     suppress the check for a minimum subject length at run time. (If it was
+     given to pcre_exec() or pcre_dfa_exec() it did work.)
+. Fixed an ASCII-dependent infelicity in pcretest that would have made it
+     fail to work when decoding hex characters in data strings in EBCDIC
+     environments.
+. It appears that in at least one Mac OS environment, the isxdigit() function
+     is implemented as a macro that evaluates to its argument more than once,
+     contravening the C 90 Standard (I haven't checked a later standard). There
+     was an instance in pcretest which caused it to go wrong when processing
+     \x{...} escapes in subject strings. The has been rewritten to avoid using
+     things like p++ in the argument of isxdigit().
+ Version 8.13 16-Aug-2011
  ------------------------
 .  The Unicode data tables have been updated to Unicode 6.0.0.
-Line 20 
 Version 8.13 30-Apr-2011
+Line 181 
 Version 8.13 30-Apr-2011
      code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
 -byte code at the end of the subject (it thought there wasn't enough data
      left).
 .  Comprehensive information about what went wrong is now returned by
      pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
      as the output vector has at least 2 elements. The offset of the start of
      the failing character and a reason code are placed in the vector.
 .  When the UTF-8 string check fails for pcre_compile(), the offset that is
      now returned is for the first byte of the failing character, instead of the
      last byte inspected. This is an incompatible change, but I hope it is small
      enough not to be a problem. It makes the returned offset consistent with
      pcre_exec() and pcre_dfa_exec().
 .  pcretest now gives a text phrase as well as the error number when
      pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
      failure, the offset and reason code are output.
 .  When \R was used with a maximizing quantifier it failed to skip backwards
      over a \r\n pair if the subsequent match failed. Instead, it just skipped
      back over a single character (\n). This seems wrong (because it treated the
      two characters as a single entity when going forwards), conflicts with the
      documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
      behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
      has been changed.
 .  Some internal refactoring has changed the processing so that the handling
      of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
      time (the PCRE_DOTALL option was changed this way some time ago: version
 .7 change 16). This has made it possible to abolish the OP_OPT op code,
      which was always a bit of a fudge. It also means that there is one less
      argument for the match() function, which reduces its stack requirements
      slightly. This change also fixes an incompatibility with Perl: the pattern
      (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
 . More internal refactoring has drastically reduced the number of recursive
      calls to match() for possessively repeated groups such as (abc)++ when
      using pcre_exec().
 . While implementing 10, a number of bugs in the handling of groups were
      discovered and fixed:
      (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
      (a|)*(?1) gave a compile-time internal error.
      ((a|)+)+  did not notice that the outer group could match an empty string.
      (^a|^)+   was not marked as anchored.
      (.*a|.*)+ was not marked as matching at start or after a newline.
 . Yet more internal refactoring has removed another argument from the match()
      function. Special calls to this function are now indicated by setting a
      value in a variable in the "match data" data block.
 . Be more explicit in pcre_study() instead of relying on "default" for
      opcodes that mean there is no starting character; this means that when new
      ones are added and accidentally left out of pcre_study(), testing should
      pick them up.
 . The -s option of pcretest has been documented for ages as being an old
      synonym of -m (show memory usage). I have changed it to mean "force study
      for every regex", that is, assume /S for every regex. This is similar to -i
      and -d etc. It's slightly incompatible, but I'm hoping nobody is still
      using it. It makes it easier to run collections of tests with and without
      study enabled, and thereby test pcre_study() more easily. All the standard
      tests are now run with and without -s (but some patterns can be marked as
      "never study" - see 20 below).
 . When (*ACCEPT) was used in a subpattern that was called recursively, the
      restoration of the capturing data to the outer values was not happening
      correctly.
 . If a recursively called subpattern ended with (*ACCEPT) and matched an
      empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
      pattern had matched an empty string, and so incorrectly returned a no
      match.
 . There was optimizing code for the last branch of non-capturing parentheses,
      and also for the obeyed branch of a conditional subexpression, which used
-     tail recursion to cut down on stack usage. Unfortunately, not that there is
+     tail recursion to cut down on stack usage. Unfortunately, now that there is
      the possibility of (*THEN) occurring in these branches, tail recursion is
      no longer possible because the return has to be checked for (*THEN). These
-     two optimizations have therefore been removed.
+     two optimizations have therefore been removed. [But see 8.20/11 above.]
 . If a pattern containing \R was studied, it was assumed that \R always
      matched two bytes, thus causing the minimum subject length to be
      incorrectly computed because \R can also match just one byte.
 . If a pattern containing (*ACCEPT) was studied, the minimum subject length
      was incorrectly computed.
-. If /S is present twice on a test pattern in pcretest input, it *disables*
+. If /S is present twice on a test pattern in pcretest input, it now
-     studying, thereby overriding the use of -s on the command line. This is
+     *disables* studying, thereby overriding the use of -s on the command line
-     necessary for one or two tests to keep the output identical in both cases.
+     (see 14 above). This is necessary for one or two tests to keep the output
+     identical in both cases.
 . When (*ACCEPT) was used in an assertion that matched an empty string and
      PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
 . When an atomic group that contained a capturing parenthesis was
      successfully matched, but the branch in which it appeared failed, the
      capturing was not being forgotten if a higher numbered group was later
      captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
      group 1 to "a", when in fact it should be unset. This applied to multi-
      branched capturing and non-capturing groups, repeated or not, and also to
-     positive assertions (capturing in negative assertions is not well defined
+     positive assertions (capturing in negative assertions does not happen
      in PCRE) and also to nested atomic groups.
 . Add the ++ qualifier feature to pcretest, to show the remainder of the
-     subject after a captured substring (to make it easier to tell which of a
+     subject after a captured substring, to make it easier to tell which of a
-     number of identical substrings has been captured).
+     number of identical substrings has been captured.
 . The way atomic groups are processed by pcre_exec() has been changed so that
      if they are repeated, backtracking one repetition now resets captured
      values correctly. For example, if ((?>(a+)b)+aabab) is matched against
      "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
      "aaa". Previously, it would have been "a". As part of this code
      refactoring, the way recursive calls are handled has also been changed.
 . If an assertion condition captured any substrings, they were not passed
      back unless some other capturing happened later. For example, if
      (?(?=(a))a) was matched against "a", no capturing was returned.
+. When studying a pattern that contained subroutine calls or assertions,
+     the code for finding the minimum length of a possible match was handling
+     direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
+     group 1 called group 2 while simultaneously a separate group 2 called group
+). A stack overflow occurred in this case. I have fixed this by limiting
+     the recursion depth to 10.
+. Updated RunTest.bat in the distribution to the version supplied by Tom
+     Fortmann. This supports explicit test numbers on the command line, and has
+     argument validation and error reporting.
+. An instance of \X with an unlimited repeat could fail if at any point the
+     first character it looked at was a mark character.
+. Some minor code refactoring concerning Unicode properties and scripts
+     should reduce the stack requirement of match() slightly.
+. Added the '=' option to pcretest to check the setting of unused capturing
+     slots at the end of the pattern, which are documented as being -1, but are
+     not included in the return count.
+. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
+     compiled something random. Now it gives a compile-time error (as does
+     Perl).
+. A *MARK encountered during the processing of a positive assertion is now
+     recorded and passed back (compatible with Perl).
+. If --only-matching or --colour was set on a pcregrep call whose pattern
+     had alternative anchored branches, the search for a second match in a line
+     was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
+     matched the line "0102" twice. The same bug affected patterns that started
+     with a backwards assertion. For example /\b01|\b02/ also matched "0102"
+     twice.
+. Previously, PCRE did not allow quantification of assertions. However, Perl
+     does, and because of capturing effects, quantifying parenthesized
+     assertions may at times be useful. Quantifiers are now allowed for
+     parenthesized assertions.
+. A minor code tidy in pcre_compile() when checking options for \R usage.
+. \g was being checked for fancy things in a character class, when it should
+     just be a literal "g".
+. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
+     appearance of a nested POSIX class supersedes an apparent external class.
+     For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
+     unescaped square brackets may also appear as part of class names. For
+     example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
+     more like Perl. (But see 8.20/1 above.)
+. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
+     was because it thought it was \N{name}, which is not supported).
+. Add minix to OS list not supporting the -S option in pcretest.
+. PCRE tries to detect cases of infinite recursion at compile time, but it
+     cannot analyze patterns in sufficient detail to catch mutual recursions
+     such as ((?1))((?2)). There is now a runtime test that gives an error if a
+     subgroup is called recursively as a subpattern for a second time at the
+     same position in the subject string. In previous releases this might have
+     been caught by the recursion limit, or it might have run out of stack.
+. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
+     happen only once. PCRE was, however incorrectly giving a compile time error
+     "recursive call could loop indefinitely" because it cannot analyze the
+     pattern in sufficient detail. The compile time test no longer happens when
+     PCRE is compiling a conditional subpattern, but actual runaway loops are
+     now caught at runtime (see 40 above).
+. It seems that Perl allows any characters other than a closing parenthesis
+     to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
+     has been changed to be the same.
+. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
+     as not to get warnings when autogen.sh is called. Also changed
+     AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
+. To help people who use pcregrep to scan files containing exceedingly long
+     lines, the following changes have been made:
+     (a) The default value of the buffer size parameter has been increased from
+K to 20K. (The actual buffer used is three times this size.)
+     (b) The default can be changed by ./configure --with-pcregrep-bufsize when
+         PCRE is built.
+     (c) A --buffer-size=n option has been added to pcregrep, to allow the size
+         to be set at run time.
+     (d) Numerical values in pcregrep options can be followed by K or M, for
+         example --buffer-size=50K.
+     (e) If a line being scanned overflows pcregrep's buffer, an error is now
+         given and the return code is set to 2.
+. Add a pointer to the latest mark to the callout data block.
+. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
+     partial match of an empty string instead of no match. This was specific to
+     the use of ".".
+. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
+     complete match instead of a partial match. This bug was dependent on both
+     the PCRE_UTF8 and PCRE_DOTALL options being set.
+. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
+     starting byte set, because \b was not being ignored.
  Version 8.12 15-Jan-2011
  ------------------------

 Legend:



Removed from v.619
 


changed lines


 
Added in v.752
 Legend:



Removed from v.619
 


changed lines


 
Added in v.752
-Removed from v.619
+Added in v.752

	ViewVC Help
Powered by ViewVC 1.1.5