Re: Proposal: Repository for fast-paced package backports
Dominik George wrote:
> Jonathan Nieder wrote:
>> 2. I am happy with the current charter of backports and I think it's
>> possible to move forward with fastpaced without having to change
>> that charter.
>
> Yep. That's exactly why the proposal changes nothing about -backports. I
> am still confused why Alex and you keep insisting that anything would be
> changing there.
It has a few points of intersection:
- Should the package begin to migrate to testing again, it must
be moved to stable-backports.
- Using the same ~bpo version namespace
- "treat it as part of backports", which I assume means that
backports users would automatically consume this repo
- new binary uploads to volatile have to undergo the
same NEW queue as backports
I don't think these are deep, inherent things (it's possible to
preserve the spirit of the proposal while removing them), but please
don't accuse me of pulling them out of thin air.
[...]
>> 3. formerer is speaking from experience when he says that it's
>> possible to make this kind of change unofficially first, learn
>> from it, and thus set the groundwork for making it official.
>>
>> If you foresee obstacles to that, can you say more about where
>> they lie? Maybe we can help address them, or maybe we can find
>> another way forward.
>>
>> If you don't see obstacles, why not start today?
>
> I think I already made those obstacles clear: Starting outside means
> buying, installing and operating at least a server vor
> volatile.debian.net (or whatever you call it), setting up and
> maintaining an upload queue, the queued, and everything around it,
> building from source for at least the most important architectures on
> hardware that needs to be there and maintained for that, etc.
Thanks. That points to who you may want to get help from:
- DSA, for hosting
- ftpmasters, in case you'd share their DAK instance
- porters, to find out what level of port + buildd support they want
to maintain
[...]
> - I do not sure I can do it right, because I do not know all the
> technical details.
That's fine. There's no time like the present to learn!
> Thus, because the change as it is proposed has such a low impact on
> anything else, I consider doing all that over again unnecessary.
>
> Don't get me wrong - I would not hesitate to go through it if it were
> for anything that could break things, or make life harder for others, or
> something like that.
I think you're underestimating the impact on other teams. That's
fine: it's probably worth it, but you will need to get buy in.
[...]
> If you know how to start with a new service at
> {volatile,fastpaced,whatever}.debian.net without having to reinvent the
> wheel for acceptign uploads, getting packages built, etc., please
> enlighten me.
backports maintainers, debian-ports maintainers, and others may have
experience with this. I don't know the best place to get advice from
them --- you may already be in the right place. :)
Sincerely,
Jonathan
Reply to: