On Tue, Nov 16, 2010 at 21:05, Barry Smith <bsmith at mcs.anl.gov> wrote:
> > I guess adding the InsertMode is fine. > > I'm not down on ever using DMComposite, I'm only down on always using > DMComposite. Okay. I think we started to discuss matrix insertion in the multi-physics context. In lieu of MatDD (or something like that) being functional, what are your feelings about a decent API for insertion? Jed -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20101116/a62cb1a2/attachment.html>
