To someone with power to push packages on Fedora 21

Zbigniew Jędrzejewski-Szmek zbyszek at in.waw.pl
Tue Oct 20 13:46:22 UTC 2015


On Tue, Oct 20, 2015 at 05:05:41AM +0200, Kevin Kofler wrote:
> Zbigniew Jędrzejewski-Szmek wrote:
> > Seriously? When I push out an update to testing, I already have done the
> > tests on it in my system, and it *think* it is correct. What would be
> > the point of pushing out something that is known to be broken?
> > 
> > The time in testing is for others to others to do the same. The result
> > is that the push to stable is based on acks from the maintainer's + n
> > independent people. You seem to be saying that the maintainer's checks
> > alone are better.
> 
> The maintainer is of course supposed to look at the testers' feedback when 
> making the decision. But it is never a good idea to blindly trust an ill-
> defined integer (the difference # positive comments - # negative comments, 
> which is only vaguely correlated to the quality of the update) going above 
> an arbitrary threshold. It makes a difference who the testers are and what 
> exactly they wrote! The software simply CANNOT make that call for you.

The testers write "works for me" or report a new bug. When things work,
there's nothing to evaluate. Doing the counting by hand is just a way to
waste time.

Zbyszek


More information about the devel mailing list