Fedora Board Recap 2009-12-17 UTC 1700

Bruno Wolff III bruno at wolff.to
Sat Dec 19 21:59:30 UTC 2009


On Sat, Dec 19, 2009 at 15:12:19 -0600,
  Bruno Wolff III <bruno at wolff.to> wrote:
> 
> There are also different causes of brokenness. Some is brokenness within
> a package or small set of packages because of sloppiness. That shouldn't
> really be happening and is definitely not desireable. Some is caused by
> a change in a subsystem used by a lot of stuff without all of the dependencies
> also getting updated. This isn't really desireable either. Some ways of
> avoiding this situation have been discussed, but there can be conflicts
> between getting the new feature in, having packagers need to jump into
> action quickly to support such changes and doing the push without having
> all of the dependencies updated.

I meant to add a bit more to this but got distracted and then sent the
message before going back to this part.

There is also another case where a new feature is being tried out and
sometimes what looks good in theory doesn't work so well in practice
even if there aren't actual bugs. I think this kind of issue is something
that rawhide is useful for testing before a bad idea makes it to an
actual release. I also think this is the kind of spear heading that Fedora
should be doing.

I don't think that being the first distro to use the new version of firefox or
whatever, really does that much to advance free software in general. I don't
think that tossing out alphas or broken betas into rawhide is really all
that useful as the upstream project can find the obvious errors without
help from Fedora.

The kernel is a bit of an exception, because getting that tested is harder
than most things and Fedora takes special steps to mitigate against bad
kernels.




More information about the fedora-advisory-board mailing list