At least the Con2PrimM routines need to be multipatch-aware.
Is this now the case?
see #676. #676 seems to be a sub-ticket of this ticket.
Yes. Some more have been added to the Zelmani version mostly since I had them added and did want to avoid extensive conflicts with Philipp once he started working on MHD and the hot EOS. This is also the reason why they are not yet in the main branch, since they touch code that keeps changing.
Mostly this affects MHD routines, *however* also eg the Marquina Riemann solver seems to be affected.
The main MHD code seems multipatch-aware however there is vector potential code that was never production ready that is not aware. I have added checks to abort the code in that case in git hash 2e99f43 "GRHydro: abort if Llama is used with AVec evolution" of einsteinevolve.