#1415: GRHydro updates ------------------------------------+--------------------------------------- Reporter: rhaas | Owner: Type: enhancement | Status: review Priority: minor | Milestone: Component: EinsteinToolkit thorn | Version: development version Resolution: | Keywords: GRHydro ------------------------------------+---------------------------------------
Comment (by rhaas):
Replying to [comment:3 knarf]:
Why is 0002 necessary? I don't like parameters with a name *_hack, for
obvious reasons. There is already an octant_hack :-) . The option prevents some aborts in prim2con which aborts for hot eos if some EOS calls failed. Essentially it affects points in the symmetry region that will be filled in by the symmetry operation later on. Affected points could be eg on the coarse grid and only be filled in by the symmetry call after restriction.
The "correct" fix to this seems to me to have a prim2confailed mask similar to con2primfailed (or re-used that mask) and only abort if the mask is set after restriction.
Until then the _hack parameters provide a workaround that lets on run some runs that otherwise fail.
I am not wedded to this parameter though (mostly I just like git and svn to be in sync as much as possible).