Firefox on Fedora: No longer funny

Thomas Spura tomspur at fedoraproject.org
Tue Oct 11 18:46:35 UTC 2011


On Tue, 11 Oct 2011 10:57:35 -0700
Adam Williamson wrote:

> On Mon, 2011-10-10 at 11:40 -0600, Tim Flink wrote:
> > On Sun, 9 Oct 2011 13:16:52 +0200
> > Thomas Spura <tomspur at fedoraproject.org> wrote:
> > 
> > > It would be great, when bodhi would allow me to add an updated
> > > mozilla-noscript to the firefox update, when I notice, that the
> > > new firefox upadate in testing breaks it.
> > > Otherwise, firefox is pushed to stable more faster, than
> > > mozilla-noscript and it's broken for some time, till it was long
> > > enought in updates-testing.
> > > 
> > > It's NOT possible to push it faster out there, because I usually
> > > don't get much karma. For the last update, I got 2 new bug
> > > reports, that noscript is broken, but not a single +1 karma for
> > > the same issue, although linking in both bug reports and being in
> > > updates-testing anyways...
> > > 
> > > (Hope to not get the usual "That's the job of AutoQA" answer...)
> > 
> > On the bright side, I don't see how AutoQA could help in this
> > situation so my answer isn't "that's the job of AutoQA". On the
> > down side, I don't really have any good answers on how to improve
> > the situation.
> 
> Well, I do. It seems pretty simple: we only have a few extensions
> packaged. We should consider extensions to be effectively API
> dependencies of Firefox, which means any Firefox update must also
> include updates for the dependencies - extensions. We should ask the
> Firefox maintainers and extension maintainers to co-ordinate so that
> each Firefox update which changes the extension API number (or
> whatever it is that causes extensions to be marked 'incompatible')
> includes rebuilds of each extension.
> 
> That way Firefox can't be pushed without the extensions. Nothing in
> the above paragraph looks particularly onerous or difficult to
> organize, to me. We're only talking about a few packages and
> packagers.

Firefox can be pushed without the extensions, because the broken deps
mail can be ignored. This won't be the case, once AutoQA is able to
stop updates, which is not a solution to the current problem, so I
didn't want to talk about that much.

Nevertheless, it could be a good first step to get notified of problems,
when firefox maintainer delay the updates and only build rawhide, so
that broken deps have a chance to notify people. That's why I'm working
towards this direction.

	-Tom


More information about the devel mailing list