Fix bugs in CaloDigitFilterTool
Found a series of bugs in CaloDigitFilterTools
- switched offset maps for SPD and noSPD - a bit surprisingly as I thought that was fixed last year. Anyway harmless as both maps are so far identical
- mixing maps between different calo when updating - most likely harmless as the tool is only use for Ecal AFAIR
- dereferencing the map pointers in case of update (?) - that makes the online monitoring to crash randomly. Hope there is/was no production with that version as I'm not sure the offsets are correctly used (applied to pile-up subtraction in CaloClusters).