Use IP for PV association of composites
Right now FDchi2 is used, which is not optimal (in particular for long lived particles).
See also @ldufour's talk https://indico.cern.ch/event/1248973/contributions/5266961/attachments/2593570/4476380/230214_IP_assoc.pdf
For HLT2 version: Rec!3460 (merged)
Edited by Maarten Van Veghel