Currently, if there is a Neumann and Dirichlet condition between two elements, the point at the element interface does not have a unique value. This PR addresses this, @vbaconnet is this an ok solution or did we come up with something else?
For meshes where the BC changes within one element (such as for boxes or the like), this is not an issue.
Before merging I guess this should be tested, but I don't have a mesh with bcs like this, does anyone have such a mesh? Preferably really small.
I need to check this is correctly done on the delta fields as well, which I guess it is not. The same thing applies to scalar and such as well, but I mostly wanted to get this discussion going.
Currently, if there is a Neumann and Dirichlet condition between two elements, the point at the element interface does not have a unique value. This PR addresses this, @vbaconnet is this an ok solution or did we come up with something else?
For meshes where the BC changes within one element (such as for boxes or the like), this is not an issue.
Before merging I guess this should be tested, but I don't have a mesh with bcs like this, does anyone have such a mesh? Preferably really small.
I need to check this is correctly done on the delta fields as well, which I guess it is not. The same thing applies to scalar and such as well, but I mostly wanted to get this discussion going.