ActsEventCnv: handle annulus surfaces correctly
I added some sanity checks to try to understand the differences between Trk and ACTS, and (with the help of @pagessin) realised that Annulus bound surfaces were not handled correctly.
These changes seem to fix these differences (many still remain, but hopefuly #65239 will help here).
BTW I will remove the sanity checks (or make them optional) once this is all working, but for the moment I think it's helpful to have them in.
Edit: I should have said that this does NOT correct the annulus cov matrices yet.
cc @ncalace
Edited by Edward Moyse