[skip-ci] clarify side effects of SetRange when fitting

[ferdymercury]

This Pull request:

Changes or fixes:

See https://root-forum.cern.ch/t/confusing-behavior-of-setrangeuser-and-fitting-range/58591

Checklist:

• [ ] tested changes locally
• [x] updated the docs (if necessary)

[rlalik]

I appreciate efforts to solve my issue, but I have an different opinion what should be the solution. I expressed it in the reply on forum: https://root-forum.cern.ch/t/confusing-behavior-of-setrangeuser-and-fitting-range/58591/5?u=rlalik

I was rather hoping on removing the side effect of TAxis::SetRange(User). There is already fit range determination by the TF1 function definition or parameters to TH1::Fit() and changing the view of the histograms (this is what SetRange does) really should not affect the fit. In my opinion the fix went in very wrong direction.