[Mulgara-dev] MGR-48 and the odd results you are seeing.

Andrae Muys andrae at netymon.com
Tue Mar 20 02:02:59 UTC 2007


Ronald,
   You say that the order of operations is:

     insert <a:foo> <a:bar> 'hi' into <local:///topazproject#foo> ;
     insert <a:foo> <a:bar> 'bye' into <local:///topazproject#foo> ;
     delete select $s <a:bar> $o from <local:///topazproject#foo>  
where $s <a:bar> $o from <local:///topazproject#foo> ;
     insert <a:foo> <a:bar> 'blah' into <local:///topazproject#foo> ;

and that the query

     select $s <a:bar> $o from <local:///topazproject#foo> where $s  
<a:bar> $o ;

returns the result

     a:foo a:bar "blah"
     a:foo a:bar "hi"
     a:foo a:bar "bye"

Can you confirm that [a:foo a:bar "blah"] was already in <...#foo>  
prior to the test (as it would have been if you ran this several  
times).  It is the only way I can explain the result you are seeing.

Either
[a:foo a:bar "blah"]  (the correct result)

or
[a:foo a:bar "hi"]
[a:foo a:bar "bye"]  (the stale result)

makes sense to me, and indeed with revision 192 I am seeing the  
latter.  If this is the case then I know what is going wrong,  
understand why your patch 'fixes' the problem, why the fix is  
brittle, and how to fix it better (I need an explicit commit at the  
end of the modify operation :).

Can you confirm that you are starting with a model containing [a:foo  
a:bar "blah"] ? If you can then MGR-48 is a quick fix.

Andrae


-- 
Andrae Muys
andrae at netymon.com
Principal Mulgara Consultant
Netymon Pty Ltd





More information about the Mulgara-dev mailing list