Non-continuous FEFunctions for high order FESpaces

[fverdugo]

@santiagobadia I have realized that high order fe spaces are not working properly. E.g.:

model = CartesianDiscreteModel(domain=(0.0,1.0,0.0,1.0), partition=(2,2))
order = 2
tags = [1,2,3,4,6,5]
fespace = ConformingFESpace(Float64,model,order,tags)
fun(x) = x[1] + x[2]
uh = interpolate(fespace,fun)
writevtk(trian,"trian", nref=4, cellfields=["uh"=>uh])

leads to a non-continuous FEFunction (here the model has 2x2 cells but I am visualizing with a finer mesh)

Everything OK for the same example with order = 1

perhaps the problem is in the generation of the cell_eqclass.

[fverdugo]

We can discuss this tomorrow (among other topics)

[fverdugo]

Finally fixed via PR #56

Closing issue.