Re: More 5 november in the release schedule
On 11/06/2016 11:59 AM, Marc Haber wrote:
> On Sun, 06 Nov 2016 09:38:00 +0000, Niels Thykier <niels@thykier.net>
> wrote:
>> Marc Haber:
>>> On Sat, 5 Nov 2016 13:46:16 +0100, Sebastiaan Couwenberg
>>> <sebastic@xs4all.nl> wrote:
>>>> [2017-Jan-05] Soft freeze (no new packages, no re-entry, 10-day
>>>> migrations)
>>>
>>> Does this really mean "once you're out, you'll stay out"?
>>>
>> Yes.
>
> That is really really bad.
Why? Any package currently in testing still has time to enter
(until roughly end of this year), so it's not like there is no
heads-up for people. And RC bugs don't lead to immediate
removal from testing, you still have quite a bit of time until
they actually cause removal of a package. That means that even
if someone reports an RC bug after the soft freeze maintainers
(and even other people, RC-bugs have a lower NMU threshold)
should have enough time to fix that before the package gets
removed. And if the problem is complicated, they have other
options: request for help on debian-devel@ and debian-mentors@,
request an exception from the release team to mark a bug as
stretch-ignore in specific cases, request an extension by the
release team to delay autoremoval so they have more time to
fix the issue, etc.
If a stable release is going to happen, there needs to be some
kind of process so that one may converge on a stable result.
What happens if you only have a single deadline to freeze
fully? Immediately before that deadline people panic because
they noticed they didn't take care of their packages enough
and upload tons of stuff on very short notice - which leads to
more bugs due to weird interactions that will then have to be
sorted out during the actual freeze. With the soft deadlines
added now, this will be relaxed quite a bit, because
everything doesn't hit at once, but it's spaced out and the
overall quality will improve.
I really don't see where you are coming from: why do you think
this makes things worse? The only people affected negatively
by this are going to be people asleep at the wheel for the
entire Stretch cycle that only wake up right before the hard
freeze. And I think that curbing that kind of maintenance
"style" is a very, very good thing.
Regards,
Christian
Reply to: