Graveyard updates

[pshriwise]

Description

These changes update some behavior of the DagMC::create/remove_graveyard function.

create_graveyard

• create_graveyard can now create a graveyard without acceleration data structures using the vertices of the model to create a bounding box as a basis for the graveyard volume.
• create_graveyard will create an implicit complement if none exists so the graveyard volume can be inserted into the model topology correctly.
• create_graveyard will only (re)construct acceleration data structures for the new graveyard volume and the implicit complement if these data structures were present when the method was initially called.

remove_graveyard

There's a bug when calling create_graveyard with overwrite == true when a graveyard is not present. Previously, this would fail when trying to remove a graveyard that isn't present. This adds a check at the beginning of remove_graveyard to simply return MB_SUCCESS if no graveyard is present.

[makeclean]

Should we ever replace an existing graveyard? Should we just not add another to the set?

[pshriwise]

Should we ever replace an existing graveyard? Should we just not add another to the set?

I can see someone wanting to replace a spherical shell graveyard in case they're learning DAGMC and didn't realize how many triangles this might add to their model without having to re-tessellate the entire thing. Granted, some kind of utility program in the repo would be better than writing it yourself own, which is what they'd have to now.

[gonuke]

I'm ready to merge but want to give @makeclean a chance to either withdraw or amplify his concern?

[gonuke]

Thanks @pshriwise