WIP: Fix/isothermal wall

Isothermal wall (WallViscous) was implemented in a non consistent way. The advection term had no BC for the energy and the the LDGNS operator used a ad hoc modification for the temperature without using the boundary condition pattern. Both issues are addressed by reading the ref temperature as a bc (similarly to the pressure). Right now lots of tests are failing due to small differences that could come from the previously lack of BC on the advection term.