[mpfa][fluxlaws] use correct flux sign switch query
The queries if the sign of a flux has to be flipped was done in the wrong way in fickslaw and fouriers law. This lead to compiler errors in the context of models that consider e.g. diffusion but no advection. Also, it would have introduced wrong flux signs in mpfa schemes in which the interaction volume regions depend on cell parameters (e.g. mpfa-l).