Re: Shouldn't we have more ftp masters ?
On Fri, Jun 02, 2006 at 11:06:29AM +1000, Anthony Towns wrote:
> On Thu, Jun 01, 2006 at 11:15:35PM +0200, Sven Luther wrote:
> > Maybe, or maybe not, but you cannot deny that the actual NEW situation is
> > satisfactory.
>
> Heh.
>
> > And actually, what i intented to say was that there where three points which
> > made me consider this a good time to post, not a proposal, but a request for
> > comment from our DPL and ftp-master about the subject.
>
> As DPL, all I've got to say is NEW policy is a matter for ftpmasters
Sure, but i guess i will ask you the same question in next years campaign :)
> to decide. As both DPL and ftpmaster, I think we could use some new
> assistants -- Mike and Randall are pretty much doing other things these
> days, and Jeroen and Ganneff have accrued more things to do over the
> past year.
Ok, but that doesn't reply to the actual question, which was how do you
reconcile your DPL plateform and the part where you advocate a more quick and
active behaviour with the inherent immobility found in the waiting-for-NEW
case.
The second question, which i guess i asked you as ftp-master and not as DPL,
was that given that :
1) as report from, i think it was, Ganneff, NEW handling of new kernel
packages was always automatic and took almost no time. I know of no case
that a decision by the kernel team was rejected by the ftp-masters, except
the experimental-in-unstable package from Bastian, but that was a new source
package.
2) i hear rumors that setting up an override for source packages is
relatively easy to do. Naturally, i cannto confirm this by myself, altough
you could.
Given that, the real question is one of confiance. Which do you believe of the
kernel team or the ftp-masters, are more competent to know what sub-package
and abi change revision are adequate ? I understand this reticence to thrust
any random maintainer, but the kernel team is handled by competent and
reasonable, and decisions are taken as a team, so i think that such reticence
in doing automated NEW for binary splits, which in the case of the kernel
packages are happenein repeatedly.
And i would like a real answer to this question, not an authoritative
non-reply like you did last time.
And even if you agree with my points, you still have the possibility to say
'but we won't implement it because ...', so please don't feel agressed by
simple discussion on the subject.
Note to other readers of this thread. I get complaints that i post too much,
and indeed it can be argued that i represent 50% of this thread, and that i
already made many replies, but what should i have done ? The original replies
did not really reply anything, so i should either let the issue fall, or
insist until i get a reply ? /me is really baffled on what the correct
behaviour is on this ... I guess with my past, i will always get the blame
whatever i do anyway though :/
Friendly,
Sven Luther
Reply to: