dolfin team mailing list archive

["Garth N. Wells" <gnw20@xxxxxxxxx>]

Anders Logg wrote:
> On Sat, Apr 25, 2009 at 08:36:06AM +0200, Ola Skavhaug wrote:
> > On Sat, Apr 25, 2009 at 7:26 AM, Garth N. Wells <gnw20@xxxxxxxxx> wrote:
> > > Ola Skavhaug wrote:
> > > > On Fri, Apr 24, 2009 at 5:04 PM, Garth N. Wells <gnw20@xxxxxxxxx> wrote:
> > > > > I've been looking at the new XML io with the intention of getting the
> > > > > plotting of MeshFunctions working again, but it's not clear to me what
> > > > > the intention is with the input of MeshFunctions. If 'intersection' is a
> > > > > MeshFunction<unsigned int>, if I do
> > > > >
> > > > >    File file("mesh_function.xml", true);
> > > > >    file << intersection;
> > > > >
> > > > > the file "mesh_function.xml" contains the XML output for a Mesh and a
> > > > > MeshFuncion (the latter being just a list of numbers). When I then try
> > > > > to read the MeshFunction,
> > > > >
> > > > >    MeshFunction<unsigned int> test_function;
> > > > >    file >> test_function;
> > > > >
> > > > > I get the error  "what():  *** Error: Mesh has not been specified,
> > > > > unable to initialize mesh function."
> > > > >
> > > > > The new XML code look sophisticated and takes some effort to understand
> > > > > - can we clear out the old code yet so there is less to trawl through?
> > > > That's a good idea. I will start on this on Monday. For the moment,
> > > > you need to do the following:
> > > >
> > > > Mesh mesh;
> > > > file >> mesh;
> > > > MeshFunction<uint> mf(mesh);
> > > > file >> mf;
> > > Is the final intention that the XML file for a MeshFunction will contain the
> > > mesh and a list of variables, or just a list of variables? If it's the
> > > latter, then FunctionPlotData could be generalised to handle MeshFunctions
> > > to faciliate plottting via viper.
> > That's a good idea. I don't like that we save both the mesh and the
> > meshfunction when writing a meshfunction to file. If we use
> > functionplotdata instead, the meshfunction and mesh is more naturally
> > stored together.
> >
> > Ola
>
> That sounds good to me.
>
> But will it then be possible to directly plot a MeshFunction stored on
> file with viper from the command line?

No.

Can someone remind me why we don't just write a .vtu file and let Viper 
plot that? Since Viper uses VTK, I would expect that to be easy.

Garth

> ------------------------------------------------------------------------
>
> _______________________________________________
> DOLFIN-dev mailing list
> DOLFIN-dev@xxxxxxxxxx
> http://www.fenics.org/mailman/listinfo/dolfin-dev