ffc team mailing list archive

[Kristian Oelgaard <k.b.oelgaard@xxxxxxxxx>]

2010/1/21 Anders Logg <logg@xxxxxxxxx>:
> On Thu, Jan 21, 2010 at 07:09:20PM +0100, Kristian Oelgaard wrote:
> > 2010/1/21 Anders Logg <logg@xxxxxxxxx>:
> > >On Thu, Jan 21, 2010 at 06:57:49PM +0100, Kristian Oelgaard wrote:
> > >>
> > >>
> > >>2010/1/21 Anders Logg <logg@xxxxxxxxx>:
> > >>>The following form files are currently failing, some because of
> > >>>missing features in the tensor representation, but here are a couple
> > >>>that I don't know how they should be handled:
> > >>>
> > >>>ElementRestriction.ufl
> > >>>*** FFC: Sorry, finite element of type "ElementRestriction" are not supported by FIAT.
> > >>
> > >>This was handled by the FFC FiniteElement class by manipulating (among other things) the entity_dofs dict.
> > >>I'll have a look once finished with quadrature and the new format.
> > >>
> > >>>QuadratureElement.ufl
> > >>>*** FFC: Sorry, finite element of type "Quadrature" are not supported by FIAT.
> > >>
> > >>I don't know if we can move this element to FIAT or if it is better to keep it in FFC. I think we just need to return a QuadratureElement object. I'll take a look at this one too.
> > >
> > >ok, good.
> > >
> > >I'm about to look more closely at supporting both quadrature and
> > >tensor representation side by side and for the same form, metadata
> > >etc.
> >
> > OK, I just need to sort out the component business for mixed elements, then the code generation for quadrature should be more or less in place.
>
> That was quick! It took me much longer to sort out the tensor
> generation.

Everything is not in place yet, I just made the simplest changes possible to make things compile again.

> Note that ufl_element.extract_component works differently from what we
> used before which was part of FFC (extract_component_element?). The
> new function returns the component relative to the nested simple sub
> element whereas the old returned an offset. If you need the offset,
> subtract what you get out of extract_component from what you send
> in...

I just use extract_component in one place, to get the sub_element corresponding to a  given component. I need this because I have to create the fiat_element to figure out what mapping is needed for that particular component, I don't use the offset for anything so I hope it just works out of the box.

Kristian

> --
> Anders
>
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.9 (GNU/Linux)
>
> iEYEARECAAYFAktYofIACgkQTuwUCDsYZdGxBgCgjnzJ8oDestrRH3yZC3wcZ8yg
> QNQAnjygLtbB51WQT8hO7ttRXQPyQRlc
> =eGio
> -----END PGP SIGNATURE-----

Attachment: signature.asc
Description: OpenPGP digital signature