Removing knot from kink makes curve rational

This occurs in 7.0.20309.6003, 11/4/2020, but not in V6:

  1. Add _Curve of any valid Degree (1-11).
  2. _InsertKink anywhere on the interior of the curve.
  3. _RemoveKnot any amount of knots from the added kink.

Per _List, curve is now rational with weights at end CPs at 1.0, but some of the other CPs have weights within 1e-12 of 1.0.

For a degree 1 curve, when only CPs at the end remain, _List reports the weights as “1”, but “is_rat = 1”.

This doesn’t appear to occur with surfaces.

Hello - I notice also that the curve rings up as a polycurve, even if when all the knots are removed it is a single span…
No idea what is possible, but I’ll add this to the heap, thanks for pointing it out.

Chuck had a look and made this YT:
RH-61777 Removing a knot from the join of a non-rational polycurve results in a rational polycurve with a single segment