dolfin team mailing list archive

[Ola Skavhaug <skavhaug@xxxxxxxxx>]

Anders Logg skrev den 29/04-2007 følgende:
> On Sat, Apr 28, 2007 at 11:46:16PM +0200, Ola Skavhaug wrote:
> > Anders Logg skrev den 28/04-2007 følgende:
> > 
> > [snip]
> > 
> > > > Thanks, that fixed the first error. The second seems to be a missing file.
> > > > 
> > > > Ola
> > > 
> > > Fixed. Sorry about this.
> > 
> > Thanks, now, pydolfin does not compile:
> > 
> > dolfin_wrap.cpp: In function 'PyObject* _wrap_new_Function__SWIG_5(PyObject*,
> > PyObject*)':
> > dolfin_wrap.cpp:16769: error: no matching function for call to
> > 'dolfin::SubFunction::SubFunction()'
> > ../../src/kernel/function/dolfin/SubFunction.h:33: note: candidates are:
> > dolfin::SubFunction::SubFunction(dolfin::DiscreteFunction*, dolfin::uint)
> > ../../src/kernel/function/dolfin/SubFunction.h:29: note:
> > dolfin::SubFunction::SubFunction(const dolfin::SubFunction&)
> > make[1]: *** [_dolfin_la-dolfin_wrap.lo] Error 1
> > 
> > Ola
> 
> It's the old touch dolfin.i problem as usual... This should be fixed
> now.

When running dolfin-stokes in taylor-hood, no file named velocity.xml is
created. Shouldn't the solutions be saved to file in xml-format?

Ola
 
> I tested two different ways of calling plot. Just calling plot(u)
> gives
> 
>   File "plot.py", line 7, in ?
>     plot(u)
>   File "/usr/lib/python2.4/site-packages/viper/viper_dolfin.py", line
>   83, in plot
>     viper = Viper(data, *args, **kwargs)
>   File "/usr/lib/python2.4/site-packages/viper/viper_dolfin.py", line
>   13, in __init__
>     self.plot(data, *args, **kwargs)
>   File "/usr/lib/python2.4/site-packages/viper/viper_dolfin.py", line
>   20, in plot
>     self.plot_function(data, *args, **kwargs)
>   File "/usr/lib/python2.4/site-packages/viper/viper_dolfin.py", line
>   60, in plot_function
>     (self.iren, self.renWin, self.ren) =
>     self.vector_plotter(self.vtkgrid, self.mesh.coordinates(), self.x,
>     self.x.min(), self.x.max(), wireframe=False)
>   File "/usr/lib/python2.4/site-packages/viper/viper.py", line 392, in
>   vector_plotter
>     vectors = self._vectors_from_numpy(coords, x)
>   File "/usr/lib/python2.4/site-packages/viper/viper.py", line 348, in
>   _vectors_from_numpy
>     directions_3d[:,:m] = directions
> 
> Adding mode="vectorfield" gives
> 
>   File "plot.py", line 8, in ?
>     plot(u, mode="vectorfield")
>   File "/usr/lib/python2.4/site-packages/viper/viper_dolfin.py", line
>   83, in plot
>     viper = Viper(data, *args, **kwargs)
>   File "/usr/lib/python2.4/site-packages/viper/viper_dolfin.py", line
>   13, in __init__
>     self.plot(data, *args, **kwargs)
>   File "/usr/lib/python2.4/site-packages/viper/viper_dolfin.py", line
>   20, in plot
>     self.plot_function(data, *args, **kwargs)
>   File "/usr/lib/python2.4/site-packages/viper/viper_dolfin.py", line
>   69, in plot_function
>     self.iren.Initialize()
> AttributeError: 'Viper' object has no attribute 'iren'
> 
> Perhaps mode="vectorfield" should be the default option for
> vector-valued functions?
> 
> By the way, mode="displacement" seems to give something reasonable for
> the Stokes demo. (It's not reasonable to plot the velocity field as a
> displacement but the output looks to be what you would expect.)
> 
> /Anders