errors and covariance with Fakes?
When I create a RooUnfoldResponse with 32 measurement and 120 truth bins using also the RooUnfoldResponse::Fake(double) function, the response looks good, but when I do the unfolding with kErrors, kCovariance or kCovToy, the message: "Error in <operator()>: Request column(120) outside matrix range of 0 - 120" is written out 32 times (i.e. once per measurement bin). And then the unfolded histogram is created, but the uncertainties are all nan values.
Is this expected behavior?
If not, or if it is not a known issue, then I can provide a working example.
Thanks a lot!