[Dune-devel] [GSoC] Update

[Oliver Sander]

Hi Xinyun,

> The updates are recorded at my blog. This is the blog address:
> https://gsoc2016xinyun.blogspot.de/.

thanks!

> Actually in my last Email I asked you something regarding the
> IntersectionIterator class? Maybe you didn't see and reply. 

Sorry for that.  Actually I just double-checked my old emails and I am not sure about
what question you mean.  Is it "Should I implement an IntersectionIterator"?
To that my answer is: only as far as the VTKWriter requires it.

> So I moved on
> with the NURBS grid manager.
>

That is a good thing to do.

> I didn't test the VTKWriter so far as I'm not very positive about what I
> have is enough for that. 

That's why I have tried to use the VTKWriter with your code: to find out whether
you have enough.

> Currently I am only able to make the codim=0 work,

What do you mean by "mak[ing] the codim=0 work"?  Does that mean that you can only
iterate over elements for know?  That's not bad: it's one of the most important
features.

> which I assume is the element dimension(Am I right?).

Codim=0 means elements, yes.

>  I don't know what to
> do with other codimensions. 

Dune grids are only required to implement iteration over elements and vertices.
Since you already have the elements, all you need to do is implement the vertex
iterators.


> And the intersection iterator is not ready.
> 

Let's wait with that until you have the VTKWriter.

> But I will try to dig into that, at least fix some bugs which can be easily
> recognizable. I will let you know if I make some progress or get stuck
> somewhere.
> 

Thanks!  And good luck!

Cheers,
Oliver

> Cheers,
> Xinyun
> 
> -----Original Message-----
> From: Oliver Sander [mailto:oliver.sander at tu-dresden.de] 
> Sent: Tuesday, August 2, 2016 3:59 PM
> To: Xinyun Li <xinyun.li at fau.de>; dune-devel at dune-project.org
> Subject: [GSoC] Update
> 
> Hi Xinyun,
> 
> I saw that you committed some more code into your repository -- nice.  On
> the other hand it has been a while since you gave us your last status
> report.  Could you please give us an update, preferably by a new blog entry?
> Thanks!
> 
> I tried to use your recent changes to replace the hand-written VTK writer by
> the standard one from dune-grid.  The small patch I used is attached below.
> With that patch the code does not compile, but the first 10 error messages
> look very easy to fix.  Can you give that a try?
> 
> Thanks,
> Oliver
> 
> 
> 
> diff --git a/src/dune-iga.cc b/src/dune-iga.cc index c816571..e3251b9 100644
> --- a/src/dune-iga.cc
> +++ b/src/dune-iga.cc
> @@ -11,6 +11,8 @@
>  #include <dune/common/exceptions.hh>
>  #include <dune/common/fvector.hh>
> 
> +#include <dune/grid/io/file/vtk/subsamplingvtkwriter.hh>
> +
>  #include <dune/iga/bsplinegrid.hh>
>  #include <dune/iga/bsplinepatch.hh>
>  #include <dune/iga/NURBSpatch.hh>
> @@ -41,7 +43,7 @@ void testBSplineGridSurface()
>    //controlNet.disp();
> 
>    IGA::BSplineGrid<dim,dimworld> grid(knotSpans, controlNet, order);
> -  const auto& gridView = grid.leafGridView();
> +  const auto gridView = grid.leafGridView();
>    const auto& indexSet = gridView.indexSet();
>    ////////////////////////////////////////////////////////////////
>    //  Write to a VTK file.
> @@ -98,6 +100,9 @@ void testBSplineGridSurface()
>      offset += ((1<<subSampling)+1) * ((1<<subSampling)+1);
>    }
>    vtkFile.write("BSplineGridTest-Surface");
> +
> +  SubsamplingVTKWriter<decltype(gridView)> 
> + vtkWriter(gridView,subSampling);  vtkWriter.write("foo");
>  }
> 
>  void testBSplineGridCurve()
> 
> 
> 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: OpenPGP digital signature
URL: <https://lists.dune-project.org/pipermail/dune-devel/attachments/20160805/334f3918/attachment.sig>