Propose fix for nc/idl map

[jkuhl-uni]

This is the fix I would propose for #283.
The interplay of this change and the lines 379, 380 is a bit strange to me.
In the proposed change, those lines would only play a role, if dtr_read < diffmeas and we need diffmeas to find the configuration numbers.
This would mean that, if I were to use dtr_read = n * diffmeas, I would end up with 1/n of the statistics and, obviously, a wrong idl, in which the indices are 1/n of their actual value).

[s-kuberski]

Hi!
Thanks for reporting the issue and for proposing a fix. I will look into both (also trying to remember if I have observed something like this in the past).