[WIP] New boundaries and floors for stability

[bprather]

Currently applies 2 fixes:

1. Applies the temperature ceiling by decreasing internal energy in fluid frame, rather than by adding mass in normal observer frame.
2. Reduces the order of reconstruction toward the poles so as to avoid relying on the ghost zones -- also avoids using any of those zones to average in case of inversion failures.

This should also come with a change defaulting electron evolution (& consequently the KTOT ceiling) to OFF.