ufl team mailing list archive

["Garth N. Wells" <gnw20@xxxxxxxxx>]

Martin Sandve Alnæs wrote:
> I'll take a look at improving the error message and checking,
> but try this (if I understand what you're trying to do correctly):
>
> kc = Function(mixed_element)
> k1, c1 = split(kc)
> dkc = TrialFunction(mixed_element)
> dk, dc = split(dkc)
>
> ...
> a1 = derivative(L1, kc, dkc)

Thanks. I managed to figure it out and it's now used in the  DOLFIN 
Cahn-Hilliard demo.

Garth

> Martin
>
>
>
> On Sat, Jun 6, 2009 at 10:34 PM, Garth N. Wells<gnw20@xxxxxxxxx> wrote:
> > I've used the UFL automatic differentiation for a primal problem, and it
> > worked great. Now I'd like to try it for the Cahn-Hilliard demo which is
> > a mixed method. I'd like to do:
> >
> >    a1 = derivative(L1, k1, dk) + derivative(L1, c1, dc)
> >
> > but I get the below error. Is there are trick for mixed elements?
> >
> > Garth
> >
> >
> >
> > Traceback (most recent call last):
> >   File "/usr/local/bin/ffc", line 186, in <module>
> >     sys.exit(main(sys.argv[1:]))
> >   File "/usr/local/bin/ffc", line 130, in main
> >     execfile(script, {})
> >   File "CahnHilliard2D.py", line 68, in <module>
> >     a1 = derivative(L1, k1, dk) + derivative(L1, c1, dc)
> >   File "/usr/local/lib/python2.6/dist-packages/ufl/formoperators.py",
> > line 142, in derivative
> >     functions, basis_functions = _handle_derivative_arguments(function,
> > basis_function)
> >   File "/usr/local/lib/python2.6/dist-packages/ufl/formoperators.py",
> > line 128, in _handle_derivative_arguments
> >     functions       = Tuple(*functions)
> > UnboundLocalError: local variable 'functions' referenced before assignment
> > _______________________________________________
> > UFL-dev mailing list
> > UFL-dev@xxxxxxxxxx
> > http://fenics.org/mailman/listinfo/ufl-dev