On Wed, 2010-09-29 at 08:11 -0400, Kamil Paral wrote:
----- "James Laska" <jlaska(a)redhat.com> wrote:
>
> Hey gang,
>
> I'm failing when looking into the list archives to review our
> previous
> conclusions on this topic. I know we talked about this before, but I
> can't find the threads.
Weird, I think we did discuss it, I even think of some ascii art showing
possible problems. But I found just this, nothing more:
https://fedorahosted.org/pipermail/autoqa-devel/2010-August/000982.html
>
> The more I think about the upgradepath test, the more I come up with
> scenarios where I (as a maintainer) would want to know the
> upgradepath
> result for my updates in 'updates-testing'. I recognize there might
> be
> some weird policy issues here, perhaps we need to get clarification
> on
> this from FESCO?
I have search through whole fedoraproject wiki and a lot of pages talk
about upgrade path, but none defines it exactly. Therefore we could ask
FESCo about exact definition, yes, that's probably a good idea (to have
some solid background for what we do).
The main question is whether we want to have upgrade path constraint
fulfilled when dealing with updates-testing repositories and what exactly
it means in this case. Does it mean upgrading from F12-updates-testing to
F13-updates? Or does it mean upgrading from F12-updates-testing to
F13-updates-testing? We don't know, and that's the problem.
That's exactly where my confusion was coming from. We can certainly
choose one path to test, but whether that is the correct specification,
I don't know.
I'll write another email talking about the problems that may
arise.
> I know there is also the potential that an update in
> 'updates-testing'
> fails during functional validation, and never lands in 'stable'.
> Isn't
> this scenario covered by also scheduling upgradepath for bodhi
> 'updates'
> requests?
>
> Sorry for coming back to this topic, I think I need some re-educating
> on
> what 'upgradepath' means.
First of all, I'd like to have my patch accepted, if possible. That should
put into a consistent state - 'updates' checked correctly,
'updates-testing'
not checked at all. Right now we are in an inconsistent state.
You are right, I corrected the original issue you noted by introducing
an odd behavior. It is running for 'updates-testing', but only checking
'stable' and 'updates' repositories for upgradepath. It seemed right at
the time, but now I realize there isn't a clear policy/expectation on
how 'updates-testing' should[n't] be included during upgradepath
calculations.
Also I have some more patches coming for upgradepath test, so pushing
into
master would save me some trouble.
When we are in a consistent state, we can unleash the debate about the
updates-testing problem :) In a following email I'll sum up the main problems
that arise and what we can do about it.
That'd be excellent, thanks Kamil for moving this forward.
Since Will is now on PTO, do you think I can push this into master,
James?
I don't see the harm in more correctly fixing the original traceback
(
https://fedorahosted.org/autoqa/ticket/226). From discussions (mail
and 1x1) I've had with everyone, it seems we all agree we need to
include 'updates-testing' in the upgradepath test, but we don't have
consensus on what the policy is. Given that, please do push your
changes into master. We'll address 'updates-testing' in your follow-on
mail.
Thanks for the clarification!
James