Fix an integer overflow in mesh_nearest_point
Description
This occurs for large meshes and when running parallel in domains. The communication uses the product np_global * grp%size, which can easily be larger than the maximum integer for large meshes. This is now replaced by using two separate MPI calls.
News snippet
Fix an integer overflow in mesh_nearest_point
Checklist
-
I have checked that my code follows the Octopus coding standards