Thanks for your help Jed. 
----- Jed Brown <jed at 59A2.org> wrote:
> On Fri, 3 Sep 2010 12:59:21 -0500, Barry Smith <bsmith at mcs.anl.gov> wrote:
> > 
> >   Jed,
> > 
> >     Thanks for your help on this.
> > 
> >     So is the current problem (with the MatLoad) essentially solved? and we 
> > can close this case.
> 
> Yes, this issue is solved until someone comes in with a mesh
> (probably/hopefully not possible with a DA) that breaks the median-of-3
> pivot selection.  (Alternatively, this basically means calling
> AOCreateBasic with a pathological ordering.)
> 
> I didn't update the Sort-With variants, I figured I would delay that
> until I have a sort that I actually like and then implement it
> generically.
> 
> Jed


Reply via email to