this merge request updates the default calibration files to "Recs2019_05_30", which include an additional systematic uncertainty extracted from the MC-to-MC comparison of rel20.7 vs rel21.
The studies are documented here: https://indico.cern.ch/event/784539/contributions/3449224/attachments/1852942/3042527/190528_r21_vs_r20p7.pdf
Several analyses (e.g. H->mm) are already using this systematic uncertainty, that was indeed needed from some time.
Best, Federico (for MCP)