On Mon, 19 May 2008, Greg DeKoenigsberg wrote:
On Mon, 19 May 2008, Rahul Sundaram wrote:
> ". If you are a Fedora fan, but would prefer your Linux desktop to have
> fewer experimental software packages, you might want to postpone the upgrade
> for a month or two, or investigate one of the distributions with a policy of
> shipping stable and well-tested software only."
>
> ... implies we don't test our packages. If you agree with analysis, fine but
> I do think our choices can be explained further.
To be perfectly honest, I *do* agree with the analysis. I think that
comprehensive testing, right now, is the biggest weakness we have. Some of
the bugs we shipped with are pretty ugly.
My current circumstance is a great example.
I'm writing this from an F9 USB key I burned using Luke's Windows USB tool on
my mother-in-law's Windows computer. Excellent!
But I'm using it because when I upgraded to F9 and then tried to use
hibernate, it bricked my laptop. Bummer!
Some of these bugs could have been found with simple test suites -- but we
haven't managed to get that work done. It's a real problem.
Side note about this... Does anyone know if we have a public list of what
our test matrix is? I'd be happy to test and sign off on some things but
I don't know what we do to test.
-Mike