yade-dev team mailing list archive

[Václav Šmilauer <eudoxos@xxxxxxxx>]

> sorry but I cannot get why it shouldn't. So if I do that (as it is yet made by the Ip2_FrictMat_FrictMat_Basic) which would be the consequence? Generally speaking, which is this design issue you mentioned?

You have to suppose some exact type of InteractionGeometry in the Ip2
functor, but you don't know anything about it (Ip2 only guarantees you
get certain InteractionPhysics types, via the dispatch mechanism).

That means, unless you use Ig2 functor producing the desired type, the
code will break (crash or whatever). The right behavior would be either
to accept any type (what we have now, at least in principle), or really
enforce InteractionGeometry type of the interation passed to that
particuler Ip2 functor.

v