Closes #473

@baperry2 Sorry, I did have that somewhere but I've got a bit too much on my plate lately.
We can move forward with the PR to have this feature available in dev, but a couple of things to keep in mind:

• currently no scalar diffusion on eb-inflow
• I'm not sure the SFINAE code in EB_BCfill.H is still useful with the PSF struct.
• keeping separate EB_inflow and EB_isoT is pretty limited. I was thinking of having a params EB2.bc_type that defaults to wall_no_slip_adiab, but can be changed to wall_no_slip_isoT (+ parsing a single value for T) or wall_slip_adiab if/when we have wall models. Or EB2.bc_type = UserDefined which trigger the use of the bcnormal_eb, bctype_eb.

Maybe those things can be adressed down the line (hopefully not too far), but I would add warning on that capability in the doc for now saying that it's still under development.

Thanks for the comments @esclapez. We do need this capability in dev for a current project, so I'll probably merge with the code as is and we can add more refinement later.

• Scalar diffusion on EB-inflow would be good to have, but most of the time advection probably dominates immediately at the inflow face so I'm fine with having it as is for the moment.
• The SFINAE stuff isn't strictly necessary but it does fail a bit more gracefully so I left it in.
• I will add some warnings in the docs and perhaps in the code that this capability is still in development and may change.