Lessons learned from !369 (tracking)
From the improvements and iterations of !369 (merged) it became clear that the current main approach for curved tracking with the analytical calculation of curved trajectories with volumes is not the path to robustness.
I suggest the following: calculate steps with whatever method, e.g. leap-frog or Boris or RK. In all cases tracks are a series of straight segments with several velocity/direction updates. Since we limit the maximal deflection to a few degrees, those straight line segments are always not very far from the actual curved track. It is numerically easy and robust to calculate the intersection of these straight segments with any volume. This will be our tracking after ICRC.