Do we need ClosestToBeam state on Downstream tracks?
Using real extrapolators in a vertex fit in the trigger is not a great idea. In run 1/2 we have also used 'interpolation' between the closest-to-beam and the first-measurement state. If we want to do that again, we need to add a ClosestToBeam state to Downstream tracks. The interpolation can be done with TrackTraj, or something equivalent.
We should then make sure the result is properly used in:
- DOCA functors (?)
- ParticleVertexFitter
- DecayTreeFit
See also: Moore!3196 (comment 7766097) !2718 (comment 5242798)
To add a ctb state to Downstream tracks one could modify this line:
To make sure that we don't compute states at nonsense z-values, we should also protect this function:
(For instance, for forward tracks it should return a z position between the z of the firstmeasurement and the rear of the velo.)