Persistency of custom objects
This is a follow up of a discussion at A&S week and of DaVinci!1186 (closed) .
Example use case:
For the matching of a charged composite particle, like
\Xi^-\to\Lambda\pi^-
, to a Velo track, we create objects that are not anticipated by the SelectivePacker (? Someone please confirm). This leads to seg-faults when trying to persist this object as output of a DecisionLine.
The current workaround is to re-run the matching from DaVinci, and work with the object directly for creating nTuples.
A change was proposed to separate Moore and DaVinci for such workflows and run the AnalysisProduction in steps. In that case, one needs to be able to persist the composite track-like particle.
/cc @msaur