Hi Pilou - there isn't a real cap there, it is a display error due to the boolean creating what is called a "non-manifold" edge in that case.

A non-manifold edge is an edge that is common to 3 or more surfaces instead being common to only 2 surfaces.

The next beta has fixed this area of booleans so that they will not create non-manifold edges anymore - the results of booleans will be broken up into only manifold pieces.

In this case, the next beta will break your object back into its original pieces though - that's because you're trying to union a piece which does not enlarge the volume of the main piece.

Your lofted surface in this case would add no thickness to the main body of the solid - to union it into the existing volume of the solid at that shared edge location, you will need to give your surface some thickness (for example by using shell), and then union it. I have attached here the result of doing that.

- Michael

Just to clarify - it is not quite that everything has to be thick, but rather that the thickness between 2 objects to be unioned at a common edge must match.

For example - this is ok, 2 objects each without thickness can be unioned or joined:




Also 2 objects that have thickness (have a volume) can be unioned:




However, a mixture of thick and thin at a common edge will not work:



This would create a situation where the bottom circular edge is connected to 3 surfaces - the bottom plane, the hole from the thick part, and the cylinder sticking out the bottom.

The next beta will separate such objects into different pieces, keeping the upper volume as one piece - the only other choice is to allow non-manifold edges, but these can create problems when sending objects to other programs, so for a while I think it is a good idea to focus on not allowing them.

- Michael

Mysteries of Nurbs disapear in the enlightening explanations :)