TriggerJobOpts: propagate LAr legacy flags to ConfigFlags in runHLT_standalone
TriggerJobOpts: propagate LAr legacy flags to ConfigFlags in runHLT_standalone Hello, When running RDOtoRDOTrigger with the legacy config, it seems that we do not set the LAr config flags correctly. In MC23, LAr flags are set by a preInclude, e.g. `--preInclude "all:Campaigns.MC23a"`, calling [Campaigns/share/MC23a.py](https://gitlab.cern.ch/atlas/athena/-/blob/main/Tools/Campaigns/share/MC23a.py), in turn calling [LArConfiguration/share/LArConfigRun3Old.py](https://gitlab.cern.ch/atlas/athena/-/blob/main/LArCalorimeter/LArConfiguration/share/LArConfigRun3Old.py). The jobproperties legacy flags are configured, but not propagated to [TriggerJobOpts/share/runHLT_standalone.py](https://gitlab.cern.ch/atlas/athena/-/blob/main/Trigger/TriggerCommon/TriggerJobOpts/share/runHLT_standalone.py) which rely on ConfigFlags. The mapping between legacy and config flags can be found in [thenaConfiguration/python/OldFlags2NewFlags.py](https://gitlab.cern.ch/atlas/athena/-/blob/main/Control/AthenaConfiguration/python/OldFlags2NewFlags.py#L109-127). After adding the mapping to runHLT_standalone, we get the following changes for MC23: * flags.LAr.ROD.nSamples is now 4 instead of 5 * flags.LAr.ROD.UseHighestGainAutoCorr is now True instead of False * flags.LAr.ROD.NumberOfCollisions is now 60 instead of 0 This bug is affecting the MC21a, MC23a and MC23c campaigns. It's impact is discussed in ATR-28590. We don't expect the counts to change in the MC CI test, because the test doesn't specify a MC campaign in the preInclude... The fix improves the agreement between CA and legacy configs (ATR-28470), the CA setup is not affected by the missing flag mapping. Cheers, Bertrand
parent
60f50874
No related branches found
No related tags found
Please register or sign in to comment