dolfin team mailing list archive

[Andre Massing <massing@xxxxxxxxx>]

Den 17.01.11 16.14, skrev Garth N. Wells:
> On 17/01/11 15:11, Andre Massing wrote:
> > Den 17.01.11 15.49, skrev Garth N. Wells:
> > > On 17/01/11 14:46, Andre Massing wrote:
> > > > Hi!
> > > >
> > > > I just realized that the previous member
> > > >
> > > > FiniteElement** coefficient_elements
> > > > and
> > > > std::vector<FiniteElement>   finite_elements
> > > > , respectively were removed from the UFC class.
> > >
> > > The coefficient finite elements are still there.
> >
> > Yes, but I need the ones which correspond to the test and trial function
> > spaces.
> >
> > > > I need these ones or
> > > > more precisely access to the underlying ufc::finite_element.
> > > >
> > > > Can anybody give me a quick hint how to achieve this in the newer DOLFIN
> > > > versions? I already scanned through the code, but could not find a
> > > > simple solution.
> > >
> > > >  From a FunctionSpace, you can extract a dolfin::FiniteElement, followed
> > > by a ufc::finite_element.
> >
> > But not from within the UFC class since the dolfin::Form form member is
> > private. Since the UFC class takes a dolfin::Form argument in the
> > constructor and a reference to the Form is stored in the UFC class,
> > would it not be meanful to make it accessible  by having a
> > const Form&  form() const;
> > member function in the UFC class?
>
> Just pass the Form and don't use UFC.
>
> The UFC class is just a convenience class for assembly, and shouldn't
> generally be used. The less that is in UFC, the easy things are to
> maintain (as I know from having made some recent changes).

Ok, I can changed that in my code, it was just convenient since I also 
need the UFCcell. To be honest, the code dependencies between UFC, 
ufc::form and dolfin::Form are sometimes a bit hard to follow and the 
overall reasoning/big picture behind the implementation is difficult to 
grasp for me :), so many thanks for the quick help!!

--
Andre

> Garth
>
> > --
> > Andre
> >
> >
> >
> >
> > > Garth
> > >
> > > > Kind regards,
> > > > Andre
> > >
> > >
> > > _______________________________________________
> > > Mailing list: https://launchpad.net/~dolfin
> > > Post to     : dolfin@xxxxxxxxxxxxxxxxxxx
> > > Unsubscribe : https://launchpad.net/~dolfin
> > > More help   : https://help.launchpad.net/ListHelp


--
André Massing       		email: massing@xxxxxxxxx
Ph.D. student			mobile: +47 469 57 401
Simula Research Laboratory
NO-1367 Fornebu, Norway