Double precision equality comparison in TH1::SameLimitsAndNBins

Hi,

In trying to run some operations on some histograms that had floating point numbers at the axis limits, I have found what looks to be dangerous double == double comparison going on here: 

 
https://root.cern.ch/doc/master/TH1_8cxx_source.html#l05482

This is causing my TH1D::Add calls to fail on histograms that for all intents and purposes should merge fine:

Error in <Merge>: Cannot merge histograms - limits are inconsistent: first: sig_bin1_SR (100, 0.000000, 1003.819791), second: sig_bin1 (100, 0.000000, 1003.819791)