port mach_jet problem from Maestro about maestroex HOT 2 CLOSED

I've looked into this again and think I've worked out what stumped me before.

• In MAESTRO, the user-specified boundary conditions are specified in mach_jet/multifab_physbc.f90. This iterates over the components, calculating different boundary conditions depending on the component.
• In MAESTROex, the physical boundaries are defined and filled here

  MAESTROeX/Source/MaestroFillData.cpp

  Lines 92 to 98 in 4dee98c

  	PhysBCFunctMaestro cphysbc(geom[lev-1],bcs,BndryFuncArray(phifill));
  	PhysBCFunctMaestro fphysbc(geom[lev ],bcs,BndryFuncArray(phifill));
  	Interpolater* mapper = &cell_cons_interp;
  	InterpFromCoarseLevel(mf, time, *cmf[0], srccomp, destcomp, ncomp, geom[lev-1], geom[lev],
  	cphysbc, 0, fphysbc, 0, refRatio(lev-1),
  	mapper, bcs, 0);

• The function that fills the boundaries is phifill, defined in bc_fill_nd.F90. This function does not take the component as an argument - it receives a 3d array representing a single component.

There does not seem to be an obvious way to make phifill aware of the index of the component. InterpFromCoarseLevel takes this as an argument, but between that and the bit where AMReX actually calls phifill I'm not entirely sure what happens to it.

It seems to have something to do with BndryFuncArray: in AMReX_PhysBCFunct.cpp, I can see that this just iterates over the components and does not pass them to the Fortran function. I think this may need to be overridden in MAESTROeX so that is possible for the Fortran function to know about the component.

from maestroex.

Fixed in #57

from maestroex.