Closed IvanYashchuk closed 1 year ago
Previously it was hardcoded that we do the conversion only with numpy.float64 dtype. Now it's just that type of the template function and provided NumPy array should match and dtype check is deferred to PyOP2.
numpy.float64
It helps with https://github.com/IvanYashchuk/fecr/issues/14, we still need CI with a build of PETSc with complex numbers support.
Previously it was hardcoded that we do the conversion only with
numpy.float64
dtype. Now it's just that type of the template function and provided NumPy array should match and dtype check is deferred to PyOP2.It helps with https://github.com/IvanYashchuk/fecr/issues/14, we still need CI with a build of PETSc with complex numbers support.