Follow-up from "RawbankView for FT/UT decoders"

The following discussion from !3225 (merged) should be addressed:

• @graven started a discussion:

  I just realised that while this is OK if this is the 'last' algorithm to add something to the RawEvent event -- if there is anything additional is added downstream of this code, then the 'map view' contained in the RawEvent will be reset, and the view becomes invalid from that point onwards (i.e. it only works if the RawEvent is immutable downstream -- but as currently other encoders still continue to update the same RawEvent, this stops working.

  And unfortunately, not only does the storage of the pointers on which relies become invalid, anything that adds a bank to RawEvent may cause existing RawBanks to be moved around in memory. So while there would be a way of making sure those pointers survive (using gaudi/Gaudi!1151 (merged)), there is no guarantee their values still point to the right memory...