Re: [Libmesh-devel] Side key computing method for Prism18

[John P. <pet...@cf...>]

On Sat, Jul 17, 2010 at 2:59 AM, Liu Kai <liu...@gm...> wrote:
> Hi,
>
>   I notice the find_neighbors() method seems not to give a correct
> answer when a Prism18 cell is adjacent to a Hex27 cell. When I trace the
> problem, I find for the side composing of same nodes, the two types
> cell's key() method override don't compute the same key value.
>
>   Hex27::key() is a re-implement overriding Hex::key() because it can
> use the quad center node to compute a unique key. But the Prism18::key()
> does not do anything special for the quad side, though they have center
> node as well. So they give difference key.
>
>   I have tried to add a re-implement version for Prism18::key() and it
> seems to work well, I put it at the tail of this mail. Am I right? Any
> suggestion or explanation is appreciated.


Hi Kai,

The code you sent looks good to me.  I am testing it now and will
check in later today.

Thank you!

-- 
John