Peter Trudelle wrote: > Gervase Markham wrote: > >> I'll come down on you because your attitude is fundamentally broken. >> Regressions are bad, yes - but you should not complain or be annoyed >> if you find them. If you want a stable build, stick with a milestone. >> If you want nightlies and the latest features, live with instability. > > > I disagree. While the nightlies are always going to be less stable than > the milestone builds, I don't think we should accept regressions and > instability, let alone live with them. It is right to be annoyed, > regressions should never get past reviews and checkin testing, but > rather than complain we should ensure the problems are well-reported in > Bugzilla.
I agree 95%. The 5% is that I think it can be acceptable to check in something which is somewhat destabilising to get more eyes on the code and testers testing it, if the code has already got as far as it can as a patch. In this case, we should accept the temporary instability as being for a greater good. But my point is that he said "I'm just so annoyed when working features get broken,", and the correct response is to not be annoyed, but to (check for duplicates and then) file bugs. Gerv
