rebdsimOptics broken on develop

[Former user]

Easy to recreate:

bdsim --file=sm.gmad --batch --ngenerate=10000
rebdsimOptics output.root output-optics.root

This is most likely introduced as a result of the perEvent histograms changes:

line 49 of rebdsimOptics shows the construction of an EventAnalysis instance, with 3rd argument true. The constructor interface was broken in 8f13ac30f7049c827b84ad8431ac14fc8d06ff27 so now this third parameter corresponds to perEntryAnalysis rather than processSamplersIn (default false). This means that rebdsimOptics is doing perEntryAnalysis and not `processSamplersIn thus segfaulting. Fixing this does fix the segfault, but now the output optics histograms are almost entirely empty.

Beyond this I cannot progress as I do not understand how SamplerAnalysis works.

[Laurie Nevay]

Fixed now in develop. As you said, the Boolean argument to EventAnalysis was for the wrong variable now in rebdsimOptics. EventAnalysis also assumed the existence of the Config singleton that doesn't exist in rebdsimOptics. This has now been fixed.

Generally in future, we'll move away from singletons as much as possible.