Normal vector problem of surface!

Hello everyone!

Is this bug?

Using the evalsrf of GH and ghpython , different results are calculated,

How to find the SUrface real normal vector with ghpython?que.gh (18.7 KB)

Meanwhile this seems to behave as expected, if you set x type hint to Brep

f = x.Faces[0]
N = f.NormalAt(y.X, y.Y)
F = f.FrameAt(y.X, y.Y)[1]

I’m curious too. Some intricacies there between GH_Surface and Surface??

Thank you! You have solved a difficult problem for me.