Fixes #803 (closed).
This introduces three changes to phantom SPH:
- The du/dt term diffusion was incorrect as it used
rho_itwice instead of both
- The dh/dt term now uses the same definition as SPHENIX
- The signal velocity did not use the new function introduced for MHD reasons.
That first point fixes the symmetry test but not sure why we did not catch it before.
The second point fixes the issue #803 (closed). However, what is strange is that we have two different definitions of this term between SPHENIX and minimal (say). The former one does not include the
f_ij terms, the latter one does. This term is not super important but it's still a weird difference.