Why does FilletEdge work in one instance, then fail in an identical other scenario?

Rhino 7.

I have a very basic closed polysurface object like this…

I was pulling my hair out in my main scene trying to run FilletEdge with .02 on it. Check what happens each time:

Notice how it breaks the object into an open polysurface with clear holes?

So, I wanted to create a clean file with just this isolated object to post here for my question, so I copied the object into a new file. I also started to screen record the problem in that new file, but lo and behold, the FilletEdge works perfectly in this separate file??

It’s exactly the same object, but in one open instance of the file it won’t fillet properly, and in the other instance it works flawlessly.

Is this likely just a bug?

Hello - the rail type where it fails is ‘DistanceFromEdge’ and where it works, ‘RollingBall’, which is generally more reliable that DistanceFromEdge.

-Pascal

In an adjacent question then, why does RollingBall fail on a ChamferEdge on the same geometry?

File…

ex.3dm (66.2 KB)

Yeh… you can see in the fillet example how complex that corner is - that stuff can be calculated for fillets but I guess not for chamfers so easily since there is nothing tangent. It is not a super easy one to fix either since the hole is not planar… if that is OK, then EdgeSrf will make something that can be joined in; if it must be planar… the answer may not be four-sided.

-Pascal

Weird
Usually RollingBall is the one that fails the most for me, and DistanceFromEdge the one that works fine.

Does it happen to have something to do with the way the surfaces are aligned ?

Hello - I do not know - I would need to have an example to even try to guess…

-Pascal