Review implicit usage of ROOT globals in TTree handling

[hegner]

There are quite a few calls to ROOT relying on global states like creating TTrees implicitly within the current gDirectory. Would it make sense to be more explicit in these calls?

[tmadlener]

I am adding the proposal of a general review of the internals of the ROOTReader here as well. Currently the handling of switching files in the events TChain is not handled properly everywhere, (see e.g. #132). it might be worthwhile to check whether it is possible to do only once per event and keep track of which "local" event is currently being read.

[tmadlener]

This is handled properly for Frame writing (introduced in #287). See, e.g.:

podio/src/ROOTFrameWriter.cc

Lines 29 to 30 in e8ae38b

	catInfo.tree = new TTree(category.c_str(), (category + " data tree").c_str());
	catInfo.tree->SetDirectory(m_file.get());

Closing this since the ROOTReader will eventually be deprecated (before v1.0)