Then in the details, I found out the 2 surfaces are Sum Surfaces. I don’t know what that is and how to handle them.
_MakeValidForV2 ‘healed’ them, and _BlendSrf works.
Using _MakeValidForV2 I have to process the surfaces 1 by 1.
Are there other means to ‘heal’ or convert the surfaces to useable geometry?
to fix it use split edges at the quads and rejoin, BlendSrf then works. the strange thing is that these cases recently show up more often. i just very recently reported one and it got confirmed as a bug, but there i produced the geometry entirely in Rhino through a simple circle that got copied resized then both circles extruded after which BlendSrf created the same twisted crap.
Sum surfaces are quite normal in NURBS: they represent a curve extruded along another curve. So most of the time when you create an extrusion in Rhino you are creating a Sum surface (there’s an exception if you use CTRL with the gumball to create an open extrusion).
If you rebuild degree 3 the maximum deviation is 7.16072e-15. So rebuild without deleting the original surfaces. Use the rebuilt surfaces to create the blend, then join the blend to the originals and hide the rebuilds.
Aha, good to know. UseExtrusions was always set to surfaces in here, but after a fresh install I forgot about it.
It would be better to always make surfaces not extrusions.
They are not very useful (to me).
Ok.
Strange that open extrusions with gumball behave different.
That’s a smart idea!
Unfortunately that doesn’t allow to change the result when using history-enabled blends.
I think the failing blends are not a sum surface problem.
Could you please take 8.30-RC1 for a spin and see if BlendSrf now does what it’s supposed to do. I made a couple of changes and while we have of course tested this internally, I’d rather discover any glitches now than when 8.30 has gone out. If 8.30 were to go out with problems, the next opportunity to fix them is in 8.31.