change of interface for the multilayer solver

[Stephane Popinet]

Dear all,

If you use the multilayer solver:

http://basilisk.fr/src/layered/hydro.h

you need to be aware of this recent patch:

http://basilisk.fr/src/?changes=20211119164142

which changes the way boundary conditions are set for vertical
viscosity: the lambda_b, u_b and dut conditions are now vector fields
rather than scalar fields, see:

http://basilisk.fr/src/layered/diffusion.h#viscous-friction-between-layers

This means that existing code needs to be modified (if it changes the
default boundary conditions). An example of what needs to be done is here:

http://basilisk.fr/src/test/wind-driven.c?changes=20211119164142

cheers,

Stephane