Re-enabled calorimeter resolution MC tests

Re-enabled calo resolution tests:

RecoConf.hlt2_reco_calo_resolution_gamma
RecoConf.hlt2_plot_calo_resolution_gamma
RecoConf.hlt2_reco_calo_resolution_pi0
RecoConf.hlt2_plot_calo_resolution_pi0

which were introduced in !386 (merged) and temporary disabled in !403 (merged) due to partial migration to new v2 CaloDigits and CaloClusters.

The updated tests benefit from the simplified format of the association tables: now there is only one table used for all clusters and hypos collections (see discussion here)

For reference, here is the output of the updated tests for gammas and pi0s:

The results are consistent with the ones from the old MR !386 (merged), but there are marginal differences w.r.t to the original tests because:

• a bug was fixed in CaloClusterReolution::operator() while the code was refactored and modernized in Lbcom!473 (merged) by @graven (many thanks Gerhard!): in the old code the last cluster was not stored in the ROOT tuple. Fortunately, the impact of this bug was totally negligible.
• NoOverlap results are now for clusters/hypos after running the CaloClusterShowerOverlap algorithm with niter=0 which still does 0th iteration. For the time being it is not possible just not to run CaloClusterShowerOverlap because it transforms clusters v2->v1; after fully migrating to v2 in the calo algortihms, the CaloClusterShowerOverlap algorithm could be just omitted for NoOverlap results.

In a few places, I've put comments in the code what should be changed after fully migrating to v2 CaloDigits and CaloClusters.

Depends on Lbcom!473 (merged) LHCb!2542 (merged)