F11 Proposal: Stabilization

[Benny Amorsen]

Jesse Keating <jkeating at redhat.com> writes:

> So again, why wouldn't people using updates-testing have caught this?
> Oh probably because the people who had systems that would have triggered
> this bug wouldn't want to use the risky repo.  Which means they would
> all fall back to the updates-tested repo you talk about and history
> would repeat itself, but maybe then you'd ask for a
> updates-tested-no-really-I-mean-it

I would be willing to run with updates-testing on by default on
several machines if I was sure that faulty packages didn't just go
away. The problem is that you're stuck with the upgraded package which
turned out to be buggy, unless you do something manually to get rid of
it. The alternative is to reissue the original package with a higher
evra.

I don't have a good solution for this problem. As it is I just test
packages which are directly relevant to my systems, and that is not
very many packages.


/Benny