Kevin Kofler wrote:
Rahul Sundaram wrote:
> That wasn't what I was talking about however. I think Josh Boyer
> understands the issue better. Essentially, the environment that spin
> maintainers do composes on should be exactly the same as the one the
> final composes are being made. Any minor change can cause issues which
> we cant easily reproduce.
But what if fixes are needed to make one of the base spins, or even another
custom spin, work?
Then, ideally both the spin owner and Fedora infrastructure folks are
aware and can coordinate on the changes required. The same concerns
apply. We should strive to minimize the differences in the environment
between test composes and the final compose.
Rahul