Problem 723

I noticed that, defining the same shape in different ways, the results depend
strongly on the
specific implementation of the geometry. What is most worrying is that, in the
case of
G4IntersectionSolid, the results depend on the order with which the intersection is
performed (solid A intersecated solid B gives a different result than solid B
intersecated
solid A). I wrote a small program to show this effect, which I placed in:

http://agata.pd.infn.it/documents/simulations/data/bugIntersection.tgz

In the package you will also find a modified version of G4IntersectionSolid
which should
solve this problem. This should work also for geant4.7.0