Assembly over interior facets is not yet working in parallel

[Simon Funke]

I am trying to write a parallel DG solver. Unfortunately, when I run the Python script with "mpirun -n 2 ... " I get following message:

*** -------------------------------------------------------------------------
*** Error: Unable to perform operation in parallel.
*** Reason: Assembly over interior facets is not yet working in parallel. Consider filing a bug report at https://bugs.launchpad.net/dolfin.
*** Where: This error was encountered inside log.cpp.
*** -------------------------------------------------------------------------

I searched the mailing list, but the last post about this issue seems to be from 2009: https://lists.launchpad.net/dolfin/msg16437.html

The following code is a reduced example of my code. It runs fine in serial and produces the above error when run with mpirun on 2 processors.

from dolfin import *

def p1dgp2(mesh):
    "Return a function space U*H on mesh from the rt0 space."
    V = VectorFunctionSpace(mesh, 'DG', 1, dim=2)
    H = FunctionSpace(mesh, 'CG', 2)
    W=V*H
    return W

def construct_div(W):
    (v, q)=TestFunctions(W)
    (u, h)=TrialFunctions(W)
    n=FacetNormal(W.mesh())
    C=inner(v,grad(h))*dx+inner(avg(v),jump(h,n))*dS
    return C

mesh = Rectangle(0, 0, 1, 1, 10, 10)
mesh.init()
W=p1dgp2(mesh)
state=Function(W)
G=construct_div(W)

tmpstate=Function(state.function_space())
rhs=action(G,state)
solve(G==rhs, tmpstate)

[Anders Logg (logg)]

This is a known bug, or rather we have not yet implemented parallel assembly for DG methods. It should come in 1.1.