Bugfixes for timeframe splitting

[nathanwbrei]

Most of these bugs were introduced during the JFactory refactoring, or the quick-and-dirty implementation of VariadicPodioOutput.

• jana:wiring_file was prefixed twice (jana:jana:wiring_file)
• VariadicPodioOutput<T> didn't support SetSubsetCollection()
• VariadicPodioOutput<T> level was incorrectly defaulted (should have been None, so that can later default to the same level as the owning JComponent)
• The PodioOutput constructor interpreted collection_name to be the short_name, not the unique_name
• VariadicPodioOutput::EulerianStore didn't replace its "transient" collections after moving them into the podio Frame.
• None of the EulerianStore() implementations set the JFactory status, leading to JFactory::Create() attempting to re-create their contents.
• JEventUnfolder::DoPreprocess() wasn't being called, leading to unfolder inputs being created inside UnfoldArrow instead of MapArrow.