@visose, thanks again for your routine, i just tested it.
Unfortunately, i found that my initial example where the curve to cap exhibits only four straight segments may have been a bit too simple.
So, while the extraction of a naked boundary curve works, the mesher won’t accept the curve, since it consists of quite a lot of straight AND curved sub-curves.
Even if the curve would get meshed properly, it would consist of a lot of triangles in order to satisfy my tolerance needs. I like your idea of employing n-gons, but i have been using these already and i think they’re limited to 12 sides, besides having problems with non-convex faces.
In my opinion, it would be better to identify the curve’s two coplanar sub-parts, close each and create a breps from closed planar loops for each sub-curve.
However, that’s not really trivial, since you’d need to generate planes for each triple of successive points, rule out co-linear triples, compare all of these planes with each other, etc. etc. Or similar.
Please let me know if you have any ideas here…
Best regards, Heinz