On Jan 29, 2014 11:13 PM, "Adam Williamson" <awilliam(a)redhat.com> wrote:
On Wed, 2014-01-29 at 16:33 -0500, Josh Boyer wrote:
> > I'd rather not confuse what is made from Fedora bits with what is
> > based on Fedora bits but includes other bits. The remix branding does
> > not seem appropriate for spins that are made purely from Fedora bits.
> That's fair. From a resource and quality perspective though, I'd
> rather not burden rel-eng and QA with having to maintain, create, and
> test spins.
The 'burden' they create on QA is precisely zero, as we explicitly do
not block releases on spins other than desktop and KDE. I don't believe
releng considers the spins much of a burden, either - it's more just
that they don't like building and pushing out stuff that no-one's even
done a sanity check on. However, we have several high quality spins that
people *do* care about and *do* test: at least the desktop spins, but I
know for e.g. finalzone puts a lot of work into the design spin.
QA does no testing of spins at all? If that's the case then I
misunderstood. If QA does test, even if they don't block the release, it
takes time and effort.
I think it's fairly presumptuous to suggest chucking all that
favour of something that doesn't even *exist* yet.
I didn't suggest chucking. I suggested moving the work to the people most
invested. Chucking would be "sorry you can use Fedora to make a different
spin" and that would be bad indeed.
Also, even if Fedora.next dies, QA had talked about lack of time to tool
and automate in general. If QA doesn't test spins today then you gain
nothing but if you do then that's at least some time back.
Spins are not free of cost. You might find it to be of little cost but
there is still cost.