Re: Bits from the Release Team - Kicking off Wheezy
On 29/04/2011 14:21, Stefano Zacchiroli wrote:
> On Thu, Apr 28, 2011 at 10:55:01PM +0000, Philipp Kern wrote:
>> On 2011-04-28, Raphael Hertzog <firstname.lastname@example.org> wrote:
>>> But I don't plan to work on any of those if the project does not agree to
>>> promote testing to something that can be advertised as usable by end-users
>>> and as something that we strive to support on a best-effort basis.
>> It's already the case that you can go and send a mail to debian-release@ that
>> package X needs to be put into testing more quickly than the set urgency or
>> asking for permission to upload to t-p-u to fix bugs if the upload is
>> entangled. And apart from that it's supported on a best-effort basis already
>> with bug fixes trickling from unstable into it and critical stuff being
>> fixed faster.
> Sure. But the procedures you mention above all require work from the
> maintainer + the release team, while they would require work only from
> the maintainer in the (hypothetical) "rolling" scenario. Under heavy
> load periods for the release team---not only freezes, but also periods
> with a huge stack of transitions to be completed---the presence of an
> extra participant might easily become a bottleneck inducing more work
> (and stress) on everybody shoulders and more delays.
In the (hypothetical) "rolling" scenario¹, one should still follow the
same procedure to get the package speedy-migrated (or uploaded to r-p-u,
if applicable). So, I don't see where you're winning. I could be missing
something here though.
¹: which has, still, to be defined clearly.
> So yes, you're right that _technically_ we can achieve the same with
> current procedures, but I argue that the underlying procedure doesn't
> scale (which is unsurprising, given it's been designed for a different
It doesn't mean that it's a problem. Well, yes, it can be enhanced… but
I'd say we didn't see periods where all Release Team members were really
that busy to not be able to add an age-days hint or approve a t-p-u
upload. I'm not sure we will see that day. And the same issue remains true
with rolling scenario (existence of that bottleneck).
Having that said, avoiding bottlenecks when possible is better.
Mehdi Dogguy مهدي الدڤي