If you happen to have opinion(s) on how dangerous (in terms of parity-violating bugs) porting the execution of the DRC rules to an engine that can attribute directional blame to specific source geometry (e.g. differentiable w.r.t. input coordinates and blaming which layer the violating source geometry is from if not blaming specific source vertices/sub-divided source line segments) will be, or whether a different open source DRC engine should be a better base to do this from, please share.
Though given the substantial hurdles that currently exist I'm gonna put the work off until early June.
