check femnl and/or cneo-elements
Observations from demo_2ds_barcneo_bending.m:
- the total Lagrange element elcneotot does not converge in the demo
- results with the `normal' updated Lagrange element elcneo depend on the number of iterations for femnl (?!)
- a sensible comparison with the demo_2ds_bar_bending seems impossible...
- results with elup only match femlin_e for nincr = 1, and (also) depend on nincr otherwise