FastJetBuilder adds redundant info?

The following discussion from !3834 (merged) should be addressed:

• @graven started a discussion:

  I'm somewhat confused about the need for addInfo -- given that the information added with jet->addInfo in addInfo is entirely computed from jet, together with (a subset of) its children, that information is implicitly stored anytime jet is persisted, as all the children from which the information is computed here are persisted. Hence the information computed in addInfo here can also be computed after reading back jet (and its children), i.e. no information is added that isn't persisted already.

assigned to @rangel and @helder

mentioned in merge request !3826

Indeed we can recalculate them when needed and don’t need to persist these quantities, but we need to have them somewhere so that we can use them, either to make a selection or to produce an ntuple. Nowadays this is done putting them in the particle extra info. In the open MR !3826, linked to MR Moore!3204, it will be done via relations tables. My intention is to have both mechanisms for a while, to allow people to adapt the codes and make tests, and in the near future drop the use o extra infos.