BufferedTree uses branch structure of first tree only
In case one processes several trees using BufferedTree/BufferedChain, the branch structure of only the first tree is used. This can result in errors like
File "/mnt/t3nfs01/data01/shome/jpata/tth/sw-nanoaod/CMSSW/python/TTH/MEAnalysis/MEAnalysis_heppy.py", line 44, in __getattr__
val = getattr(self.__dict__["tree"], attr)
AttributeError: 'TChain' object has no attribute 'HLT_Mu8_TrkIsoVVL_Ele23_CaloIdL_TrackIdL_IsoVL'
in case a following tree does not have a branch that existed in the first tree. Need to make the branch buffers update themselves on Chain tree update.