It might be a good idea, to detect all horizontal surface triangles, sort them for size, use those relevant triangles to break the model into z-parts and adapt the layer heights for each z-part separately. Therewith horizontal surfaces may be printed at exact z-position as needed for constructive models. The current algorithm leads to inacurate z-positions in the extent of the minimal layer height set.
Is anybody able to implement this idea in the relevant c++ code of cura-engine?
It is to expansive for me to implement the idea due to other work at the moment.