Closed AnderGray closed 5 months ago
Hey, thanks for reporting this. I think it would be good to potential move away from these compiled solvers. I see two potential solutions:
radius.sh
on unix
machines and radius.ps1
on windows machines.Any preferences?
I find the 100% Julia solution to be the cleanest, unless it defeats the purpose of the external model test?
As long as I call it as a separate process it should't be a problem. I'll see what can be done.
Package was failing when testing on my HPC machine, and getting a confusing error from Julia.
Eventually found it was the c++ external solvers used in some of the tests
i.e this one: https://github.com/FriesischScott/UncertaintyQuantification.jl/blob/e62374b006d449b8e8b5a0a4aa0fc4a356d6da57/test/solvers/src/radius.c#L1
Perhaps it's worth including little build script for these c++ external solvers, and printing a custom error message in the case that these tests fail.