bug in calorimeter hit positions

[gaede]

The global position of calorimeter hits is off by ~half a cell size.

See attached plots that shows the distance in x-y of the global step position as given by Geant4 and the global hit position that we create in Geant4SensitiveAction<Geant4Calorimeter>

ddsim_calo_hits_position.pdf

• the positions are systematically wrong by [0,1] cell dimension (5*mm) most of the time ( expected would be [0.,0.5] cell dimensions)

ddsim_calo_hits_position_logy.pdf

• there are spurious hits where the difference is a few cell dimensions

• plots are created with this debug code: https://github.com/gaede/DD4hep/blob/debug_cell_positions/DDG4/plugins/Geant4SDActions.cpp#L145-L150 and particle gun (10 GeV photons, ILD Ecal (5mm*5mm cells)).

[danieljeans]

I may have an explanation and fix; confirming...

[andresailer]

Fixed by #227 ?

[danieljeans]

yes, this should be fixed