Periodic refreshing of histogram files
I am working on software to perform online monitoring of data from our DAQ through a data dispatcher. This job will connect to the dispatcher stream and process events indefinitely (ideally if the job never crashed it could be started once in the lifetime of the experiment). However, I would like this job to periodically close its histogram file, reset all the histograms in memory (although could easily be accomplished in user code), and open a new file. This would occur, say, at subrun boundaries. As I understand it, in the current framework, the histogram files are only closed at the endJob stage and there is no provision for starting a fresh file.