On Wed, 2004-10-27 at 15:24, ole.rohne@cern.ch wrote:
> Jim> I have spent some time with your patches. At first, I brazenly assumed
> Jim> that I understood them well enough to apply them to the current cvs. I
> Jim> found that make check hung on rtest2.mac. The lines
>
> Jim> M:MATRIX([A,0],[B,1]);
> Jim> M^2;
>
> Jim> send maxima into an infinite loop.
>
> The culprit was NLEFT in clmacs.lisp - it exposes a very subtle
> difference in the LOOP/SLOOP semantics. I accidentally overlooked it
> when I ported my patches to recent-CVS, and fooled myself at running
> the tests with the stock maxima.core - no wonder I never saw a timing
> difference...
>
> I apologize for your wasted time, and I hope you could have another
> go with the appended patches.
The new patches look good. I think they are ready to go in. Thanks for
your efforts.
--Jim