Absence of detectors should not cause EICrecon to stop

[DraTeots]

User story: tracking experts are annoyed that simulation of full detector takes too long and switch off either the most simulation-long calorimeters or use tracking_only.xml setups. When the reconstruction is run on the same detector setup EICrecon fails when calorimetry reconstruction try getting nonexistent geometry via decoders. There are workarounds, but in general this should be fixed.

Workaround 1: (this one is tested) use -Ppodio:output_include_collections to override reconstruction collections and triggered factories to switch off calorimetry reconstruction:

-Ppodio:output_include_collections="MCParticles,ReconstructedChargedParticles,GeneratedParticles"

Workaround 2(general): (in theory) if one uses full detector setup on files that are simulated with partial detector setup, it probably should be fine as there are just no hits for missing part. But fun sideeffects could arise.

[DraTeots]

related #300