Merge branch 'feature/exact-solution-from-ufl' into 'master'

Feature/exact solution from ufl

Adds three form compiler arguments that:

- Tell the form compiler which expression of the ufl file represents the exact solution.
- Enable comparison dof vectors of numerical solution and interpolation of exact solution.
- Enable calculation of l2errorsquared of difference between numerical solution and interpolation of exact solution.

These methods only work for scalar equations. For Mixed-/Vector-/Tensorelements an exception is raised.



See merge request !7