On Sun, Dec 18, 2016 at 03:21:14PM +0100, Yves-Alexis Perez wrote: > But at some point it's just not practical.
This is becoming interesting. So, would you be in favour of establishing some kind of "probability of failure" threshold under which every FTBFS bug stop being severity:serious? [ In some sense, you have already done that in this report ]. I asked the release managers what to do with packages which FTBFS randomly, but they didn't give me any useful guideline yet. Based on your "at some point" above, do you personally think that probability-based RC-ness is the way to go, and if yes, where would you put the threshold? Thanks.