ffc team mailing list archive

[Anders Logg <logg@xxxxxxxxx>]

Nice, I will update the references.

--
Anders


On Tue, Jan 26, 2010 at 12:32:17PM -0000, noreply@xxxxxxxxxxxxx wrote:
> ------------------------------------------------------------
> revno: 1576
> committer: Marie E. Rognes <meg@xxxxxxxxx>
> branch nick: ffc-unstable
> timestamp: Tue 2010-01-26 13:28:46 +0100
> message:
>   Updated ufctest to give more correct values for
>   interpolate_vertex_values. Could some update references? (Marie is a
>   bit unsure about her ufl/ffc setup.)
>
>   It seems that now only MixedPoisson fails for
>   interpolate_vertex_values (which it should because the old one is
>   wrong.)
> modified:
>   test/regression/ufctest.h
>
>

> === modified file 'test/regression/ufctest.h'
> --- test/regression/ufctest.h	2010-01-26 09:10:52 +0000
> +++ test/regression/ufctest.h	2010-01-26 12:28:46 +0000
> @@ -268,15 +268,19 @@
>    }
>
>    // evaluate_dof
> +  double* copies = new double[element.space_dimension()*value_size*derivative_size];
>    for (uint i = 0; i < element.space_dimension(); i++)
> -    print_scalar("evaluate_dof", element.evaluate_dof(i, f, c), i);
> +  {
> +    copies[i] = element.evaluate_dof(i, f, c);
> +    print_scalar("evaluate_dof", copies[i], i);
> +  }
>
>    // evaluate_dofs
>    element.evaluate_dofs(values, f, c);
>    print_array("evaluate_dofs", element.space_dimension(), values);
>
>    // interpolate_vertex_values
> -  element.interpolate_vertex_values(vertex_values, values, c);
> +  element.interpolate_vertex_values(vertex_values, copies, c);
>    print_array("interpolate_vertex_values", (c.topological_dimension + 1)*value_size, vertex_values);
>
>    // num_sub_dof_elements
>

Attachment: signature.asc
Description: Digital signature