Different result when using fenics_mixed_dimensional branch

Errors
1

Hi, I am using “fenics_mixed_dimensional” branch downloaded from Docker Hub.

To test the use of mixed dimension works, I ran the following simple script:

import numpy as np
import dolfin as dl


class BottomBoundary(dl.SubDomain):
    def inside(self, x, on_boundary):
        return dl.near(x[1], 0) and on_boundary


nx = 1
nz = 1

mesh = dl.UnitSquareMesh(nx, nz)
bottom_bd = BottomBoundary()

marker = dl.MeshFunction("size_t", mesh, mesh.topology().dim() - 1, 0)
bottom_bd.mark(marker, 1)
submesh = dl.MeshView.create(marker, 1)

Vh0 = dl.FunctionSpace(mesh, "Lagrange", 1)
Vh1 = dl.FunctionSpace(submesh, "Lagrange", 1)

v0 = dl.Function(Vh0)
m0 = dl.Function(Vh0)
m1 = dl.Function(Vh1)

v0.vector()[:] = np.random.rand(Vh0.dim())
m0.vector()[:] = np.ones(Vh0.dim())
m1.vector()[:] = np.ones(Vh1.dim())

ds = dl.Measure("ds", domain=mesh, subdomain_data=marker)
f0 = v0 * m0 * ds(1)

ds1 = dl.Measure("dx", domain=submesh)
f1 = v0 * m1 * ds1

print(dl.assemble(f0))
print(dl.assemble_mixed(f1))

I expect the printed results should be same, but they are different.

Could you please help me out of this issue? Thank you!

2

Hi,
Indeed the assembly of f0 and f1 should give the same result! I was able to reproduce your problem with the latest Docker, and I noticed that older Docker containers were giving the expected result.
I pushed an updated Docker container with a fix and I added your MWE as a test to make sure this keep giving the expected result;)