Rhino 4 running on Windows 7
Strange bump from sweep-2-rails:
The file sweep2 bump.3dm has two very similar surfaces, both produced by sweeping the same rails and the same cross-section curves.
The one on the left has a strange (and undesired) bump just before the corner, at the top of the long section. The one at the right does not have this bump.
The rails and cross-section curves are (copied) to the left of both surfaces. The cross section curves are the curves parallel to the line x=y=0. The lower rails are at z=0; the upper rails are at about z=4 (they vary in height).
The left surface, with the bump, is the result of one sweep2, chaining the two sections of one rail, and then the two sections of the other rail, and then selecting all five cross section curves.
The right surface, without the bump, is the result of two separate applications of sweep2. The first sweep2 uses the long rails, and four cross section curves. The second sweep2 uses the short rails and two cross section curves. Then one joins the two resulting surfaces.
Yes, I have gotten what I wanted. However, it took a couple of hours to disentangle this. I (and I am sure many Rhino users) would much rather spend my time doing design, rather dealing with idiosyncrasies in Rhino.
McNeel Associates should either adjust the sweep2 function to get rid of such idiosyncrasies, or at least put a statement in the documentation of the sweep2 of how to avoid the conditions under which these idiosyncrasies occur.
I do many sweeps and lofts and it is not weird to me that at a knuckle of rails the result of sweep may not be smooth. Of course, your second approach (two separate sweeps, each on rails without knuckles) is correct and, I think, the only appropriate.