CellSize(mesh) looks wrong in 1D

[Garth Wells]

I am writing some testing code for stabilisation methods (SU and the like) and need the size of the cells therefore. I thought in 1D the CellSize would be the size of the elements, i.e. 1/N if I have N equally sized elements; the same as the CellVolume. However, there is a factor of 2 between those and I don't know why. The documentation is unfortunately rather "sparse" on this issue. Can you please explain to me what CellSize exactly returns? Also in higher dimensions? Thanks a lot.

Here is some sample code, showing the issue:

In [10]: mesh=UnitInterval(10)
*** -------------------------------------------------------------------------
*** Warning: UnitInterval has been deprecated in DOLFIN version 1.1.0.
*** The class UnitInterval has been replaced by UnitIntervalMesh.
*** -------------------------------------------------------------------------

In [11]: assemble(CellVolume(mesh)*dx, mesh=mesh)
Out[11]: 0.1

In [12]: assemble(CellSize(mesh)*dx, mesh=mesh)
Out[12]: 0.2

[Garth Wells (garth-wells)]

Looks like a problem in ffc/codesnippets.py - should circumradius be divided by 2?