This should hopefully fix the CI failures caused by the typo fix in !53398 (merged).
I'm going to enable full unit tests on this one to avoid another round of 'whack-a-mole' hopefully.
The current situation is that the TileSamplingFraction conditions are being taken from a file for digi/MC reco now I think. The G4Version was being correctly determined from the input file metadata, but the version from the conditions file seems to be more like the sampling fraction than a Geant4 version.
Geant4 simulation version (10.01) is not compatible with version used to calculate Tile SF (0.34)In this MR, I've suppressed the comparison of the conditions G4Version in this case again. The output doesn't change, so this must reproduce the previous behaviour.
It would be nicer if digi/reco could take the sampling fraction conditions from COOL, as I had originally intended in !52397 (merged), but this seemed to cause problems for a number of unit tests in the TileByteStream and TileRecUtils packages.
It probably needs someone more familiar with the TileCalorimeter code to make this next step (assuming it makes sense). @harkusha could you comment please?