With this PR, #574 and https://github.com/arborx/ArborX/pull/353 All the tests pass. I had to relax the tolerance for two tests. One is interpolation using H_DIV which nobody is using. The other is a MLS test which was already a little bit strange. All the other tests were using a tolerance of 1e-14 but this one was using 1e-11.
With this PR, #574 and https://github.com/arborx/ArborX/pull/353 All the tests pass. I had to relax the tolerance for two tests. One is interpolation using
H_DIV
which nobody is using. The other is a MLS test which was already a little bit strange. All the other tests were using a tolerance of 1e-14 but this one was using 1e-11.