AlgoOffline
already depends on FairLogger
3 lines above, not sure about the outcome, it may explain some of the build warnings and the lower than I expected reduction in the weekly builds log sizefles_logging
header log.hpp
was used instead of the compatibility one, with sometimes a mix in the same folderNO_ROOT
flag was propagated to multiple higher level reco library (steer, offline steer, probably also analysis ones...) through Littrack
which was depending on Littrackparallel
which itself was wrongly depending on the online KfCore
instead of KfCoreOffline
. This combined with the includes in the Digi
classes to force a dependency on fles_logging
in offline libraries which should have only depended on FairLogger
Difference between the last two versions shows how involved the debugging process was (around 10 hours on my side)
Redmine: Refs #3399