← Back to team overview

ufl team mailing list archive

Re: [HG UFL] Fixed lhs/rhs bug, turns out it was in the list_tensor handler.

 

I fixed _something_...

Now I'm getting:


Visiting Zero: Zero((), (), {})
Zero is not supported (yet).

*** FFC: Zero is not supported (yet).
*** FFC: To get more information about this error, rerun FFC with --debug.


And sfc fails to compile the form for some other reason.

Martin



On Tue, Sep 22, 2009 at 9:26 PM, Martin Sandve Alnæs <martinal@xxxxxxxxx> wrote:
> Further simplification that still reproduces the problem:
>
>
> cell = triangle
> d = cell.d
> I = Identity(d)
>
> vector = VectorElement("Lagrange", cell, 2)
> scalar = FiniteElement("Lagrange", cell, 1)
> element = vector + scalar
>
> v, q = TestFunctions(element)
> u, p = TrialFunctions(element)
>
> F = inner(grad(v), grad(u) - p*I)*dx
>
> #a = lhs(F)
> L = rhs(F)
>
>
> Note that you didn't use a mixed element.
>
> Martin
>
>
>
> On Tue, Sep 22, 2009 at 9:16 PM, Harish Narayanan
> <harish.mlists@xxxxxxxxx> wrote:
>> Martin Sandve Alnæs wrote:
>>> I agree that the error is completely incomprehensible,
>>> but in sigma(u, p0) you're adding terms with a trialfunction
>>> only in one of the terms. Can you try with a valid form?
>>
>> I was overzealous in reducing the size of the form. As you noted in your
>> next e-mail, changing p0 to p results in the same error.
>>
>> Harish
>>
>


Follow ups

References