volume/radius giving wrong results?

[Marie Rognes]

I'm puzzled: I would expect

from dolfin import *
mesh = UnitIntervalMesh(3)
volume = mesh.ufl_cell().volume
radius = mesh.ufl_cell().circumradius
print "volume = ", assemble(volume*dx, mesh=mesh)
print "radius = ", assemble(radius*dx, mesh=mesh)

to return 1.0 (in both cases). I do however get 1.0/3.

Can anyone reproduce? If yes, is my expectation wrong or
is really volume/circumradius wrong in FFC?

[Martin Sandve Alnæs (martinal)]

The volume of each cell is 1/3, and integrating 1/3 over the unit interval yields 1/3. Sounds about right to me.

[Kristian B. Ølgaard (k.b.oelgaard)]

Yes, I can confirm this as I spent a day hunting down this 'bug' :)

Kristian

On 11 December 2012 15:38, Martin Sandve Alnæs
<email address hidden> wrote:
> The volume of each cell is 1/3, and integrating 1/3 over the unit
> interval yields 1/3. Sounds about right to me.
>
> --
> You received this bug notification because you are a member of FFC Core
> Team, which is subscribed to FFC.
> https://bugs.launchpad.net/bugs/1088918
>
> Title:
> volume/radius giving wrong results?
>
> Status in FEniCS Form Compiler:
> New
>
> Bug description:
> I'm puzzled: I would expect
>
> from dolfin import *
> mesh = UnitIntervalMesh(3)
> volume = mesh.ufl_cell().volume
> radius = mesh.ufl_cell().circumradius
> print "volume = ", assemble(volume*dx, mesh=mesh)
> print "radius = ", assemble(radius*dx, mesh=mesh)
>
> to return 1.0 (in both cases). I do however get 1.0/3.
>
> Can anyone reproduce? If yes, is my expectation wrong or
> is really volume/circumradius wrong in FFC?
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/ffc/+bug/1088918/+subscriptions

[Marie Rognes (meg-simula)]

Aaah... , thanks!