Handling error raw banks inside HLT1

This branch is completely based on fest_rebase_master branch (!650 (merged)) which has to be merged first. Test were run on MooreOnline environment, together with MooreOnline!70 (closed) with samples coming from the October pilot beam test /scratch/eb_testing/lhcb_bu_*.mep (inside online servers).

While reading a MEP file, the events are handled into slices. At runtime, we added a vector to the slice object which has information about subdetectors (info from the source id), the events range handled by the slice and the lhcb raw bank numbers (collected for each single MFP). Once this vector stored into the slice, it can be passed to a BankAndOffsets object which is accessible inside algorithms from runtimeoptions.

A HOST algorithm HostErrorBanksCut read the vector from the BankAndOffsets object and check which lhcb raw banks are inside the MEP. If an error bank is found the event is flagged (at the moment we created an HLT1 line that skips events with at least one error bank).

This implementation allows to distinguish also from which subdetector the error bank is coming from (which could be useful for example for some monitoring jobs).

Goes together with MooreOnline!70 (closed)

It should go also with LHCb!3435 (merged)