DQ has not been proven to work. But keep DQ degree 0 running in order to support constants.
Also note that it is not known whether the problem is in DQ or ds(quadmesh), dS(quadmesh). So eventually it is facet integrals on Q/DQ which might need to be disabled rather than DQ. Could you try to test @jackhale?
The consequences of this change have not been tested with the DOLFIN suite yet. It merely keeps neumann-poisson demo running and fails on DQ.
DQ
has not been proven to work. But keepDQ
degree 0 running in order to support constants.Also note that it is not known whether the problem is in
DQ
ords(quadmesh)
,dS(quadmesh)
. So eventually it is facet integrals onQ/DQ
which might need to be disabled rather thanDQ
. Could you try to test @jackhale?The consequences of this change have not been tested with the DOLFIN suite yet. It merely keeps
neumann-poisson
demo running and fails onDQ
.