... although apparently .subs doesn't work for TestFunction or TrialFunction ...

Had to use u[0,:] for that.
From: Cotter, Colin J <colin.cotter@imperial.ac.uk>
Sent: 23 November 2019 14:17
To: firedrake <firedrake@imperial.ac.uk>
Subject: Re: [firedrake] VFS
 
Thanks Lawrence,
  Good way to check - picking the dim of the VFS different from the gdim.
I had also forgotten about .sub, useful.

all the best
--cjc
From: Lawrence Mitchell <wence@gmx.li>
Sent: 23 November 2019 14:15
To: Cotter, Colin J <colin.cotter@imperial.ac.uk>
Cc: firedrake <firedrake@imperial.ac.uk>
Subject: Re: [firedrake] VFS
 


> On 23 Nov 2019, at 14:02, Cotter, Colin J <colin.cotter@imperial.ac.uk> wrote:
>
> VectorFunctionSpace(mesh, "BDM", 1)
> ?
>
> Is the first index the dim copies and the second index the vector component of the copies of the BDM spaces?

Yes

In [3]: V = VectorFunctionSpace(mesh, "BDM", 2, dim=4)                                              

In [4]: f = Function(V)                                                                             

In [5]: f.ufl_shape                                                                                 
Out[5]: (4, 2)

In [6]: f.sub(0)                                                                                    
Out[6]: Coefficient(WithGeometry(ComponentProxyFunctionSpace(<firedrake.mesh.MeshTopology object at 0x1202e31d0>, FiniteElement('Brezzi-Douglas-Marini', triangle, 2), name=None, index=None, component=0), Mesh(VectorElement(FiniteElement('Lagrange', triangle, 1), dim=2), 1)), 7)

In [7]: f.sub(0).ufl_shape                                                                          
Out[7]: (2,)

In [8]: f.sub(1).ufl_shape                                                                          
Out[8]: (2,)

In [9]: f.sub(2).ufl_shape                                                                          
Out[9]: (2,)