ffc team mailing list archive

[Anders Logg <logg@xxxxxxxxx>]

On Wed, Sep 09, 2009 at 07:29:05PM +0200, Martin Sandve Alnæs wrote:
> On Wed, Aug 12, 2009 at 3:08 PM, Harish
> Narayanan<harish.mlists@xxxxxxxxx> wrote:
> > I'm running into an error when I attempt to use lhs() and rhs() with
> > mixed elements. I have attached two small .ufl files, one for the
> > Poisson equation (which works) and the other for the Stokes equations
> > (which fails with the following error).
> >
> > Preprocessing form file: stokes-broken.ufl --> stokes-broken.py
>
>
>
> > Wrapping non-UFL expression. This is experimental and will likely break
> > many algorithms!
>
> By the way, I recommend removing this experimental feature.
> It hides the proper errors. Anders asked for it, but I don't think it's
> used anywhere, and I don't believe it can be properly handled
> in most algorithms.
>
> Martin

I don't even remember what this feature is. Is it intended for things
like

  L = v*sin(x)*dx

?

--
Anders

Attachment: signature.asc
Description: Digital signature