I've added a couple of options to have a smooth interface between the sphere and ambient medium in the collapse and photon tests. Another added option is constant pressure, given a density profile, where the pressure in the sphere equals the pressure at the outer sphere radius. I've added the new parameters to the docs and it passed the CollapseTestNonCosmological test.
This looks like a useful addition. My only concern is that the spherical CollapseTest is already extremely complicated, with many knobs and switches that have been added over the years and that it might be a good idea to come back and try to streamline a bit. That being said, I don't think my concern should hold up acceptance.
I completely agree about both points. My hope is that we can either use yt => GDF => Enzo, or possibly implement yt-like operators and fluid operations that can do things like this. But, this by itself seems great.