On Nov 9, 2011, at 4:55 PM, Jed Brown wrote:

> On Wed, Nov 9, 2011 at 16:50, Barry Smith <bsmith at mcs.anl.gov> wrote:
> We don't usually do things that way hence I think a specific checking 
> function is better.
> 
>   3 seconds later I like the idea of returning null if it does not exist and 
> not having the other checking function. Mark me down for the returning null 
> approach.
> 
> So the concern is what should VecGhostUpdateBegin/End do for non-ghosted 
> Vecs? In the serial case, there is no harm in doing nothing because in the 
> usual situation, there are just zero ghosted entries. For user-defined Vecs, 
> what should happen?

   Error in parallel not sequential.

> 
> Should VecGhostUpdateBegin just fail for anything that is not explicitly 
> ghosted? I don't have a problem using a Vec_MPI in serial so that the 
> presence of exactly 0 ghosts can be explicit.


Reply via email to