Maude-help mailing list

[Jon Haugsand <jonhaug AT ifi.uio.no>]

* Steven Eker
> Hi,
> 
> Thanks very much for the bug report. There is a serious and subtle bug in 
> the 
> sort computation code. Specifically, when an A or AU argument list becomes 
> long enough, Maude switches to a persistent deque representation which 
> allows 
> certain kinds of patterns to be matched in constant time, certain kinds of 
> rhs replacements to be reflattened in constant time and sort computations 
> to 
> be performed incrementally in constant time. There is a bug in the pointer 
> reversal code for the latter algorithm such that sorts previously 
> calculated 
> for parts of an argument list are not used in the final calculation, with 
> the 
> possibility that the wrong final sort will be calculated. Since the 
> persistent deque representation is highly redundant you have to do just the 
> right sequence of rewrites to hit the bad case.

Ah.  Cannot say I got this all through, but never mind.

> 
> The will be an alpha release with a fix in the next day or so, with a 
> public 
> release in a week or two. If you would like to be an alpha tester let me 
> know 
> and I will send you the alpha release site details.

I think I'll wait for the public release.  But thanks.

-- 
Jon Haugsand
  Dept. of Informatics, Univ. of Oslo, Norway, 
mailto:jonhaug AT ifi.uio.no
  http://www.ifi.uio.no/~jonhaug/, Phone: +47 22 85 24 92