From: Roy S. <roy...@ic...> - 2011-04-20 18:57:41
|
On Wed, 20 Apr 2011, Kirk, Benjamin (JSC-EG311) wrote: > On 4/20/11 1:49 PM, "Roy Stogner" <roy...@ic...> wrote: > >> I'll fix this now. Any FIN-S repercussions I should know about? > > Nope. That's a relief. I don't see us using a PointLocator or MeshFunction on the BoundaryMesh, but I figured the mostly likely way for you to run across this bug would be to get bitten by it... > I'll check CATPISS because it uses this to extract TC traces > in-depth. The current bug was probably causing us to use an OctTree when a > QuadTree would do, which is an efficiency thing, not an error. That's fine; what concerns me is that it could also potentially cause us to use a QuadTree when an OctTree was necessary. Not a fun thing to realize when working on a 3D code that does a few fancy BoundaryMesh tricks. --- Roy |