Bug in RooExtendPdf?

It looks almost certainly as a typo-like bug in the code to me, but since I may at times confuse things, I thought I’d ask before opening a report:

RooExtendPdf can optionally use a sub-range of observable to calculate expected number of events.
Say I have a pdf depending only on e=RooRealVar(“e”,“energy”,0,10) that has a subrange “signal”=(1,8). I would assume the ::expectedEvents(e) using the sub-range option should be smaller. But in fact, it’s larger.
I see in the code the fractional integral of the pdf in the signal range is calculated correctly, but then the total number of events is divided by the fraction, instead of multiplied.

Am I misunderstanding the basic idea, or is it, indeed, a typo?