[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#333858: Proposed improved patch for yaird



On Mon, Oct 17, 2005 at 09:57:49AM +0200, Jonas Smedegaard wrote:
> On Sun, 16 Oct 2005 23:57:59 +0200 Sven Luther <sven.luther@wanadoo.fr> wrote:
> > On Sun, Oct 16, 2005 at 10:35:45PM +0200, Jonas Smedegaard wrote:
> > > But please wait doing too much about that. I'd like to have it setup
> > > differently - or at least discuss having setup differently.
>                               ^^^^^^^

> Erik: I see 3 sane scenarios tied to using Alioth SVN:
> 
> 1) You join the Debian-kernel team and we maintain upstream yaird and
> debian subdir together. Every upstream release - even if only related
> to the redhat - is also a Debian release.
> 
> 2) You join the Debian-kernel team and we maintain upstream yaird and
> debian subdir separately within the project. Upstream releases and
> Debian releases are separate.
> 
> 3) You join the Debian-kernel team if you want to work directly on the
> debian subdir of yaird - or coordinate with us as now and concentrate
> on the code itself - maintained as a separate Alioth project or
> elsewere.
> 
> 4) Yaird is setup as a separate Alioth project - either all of it (as
> in 1 or 2) or only the debian subdir (as in 3).
> 
> 
> One reason for 3 or 4 is the amount of "noise" if you are not
> interested in what else goes on regarding kernel maintainance in
> Debian. If you are in for some information overload, I suggest 2 - and
> hope that in addition to more convenient maintainence of yaird compared
> to now (which is somewhat clunky, as we've both agreed on) you also
> gain interest in the other work here and apply you insight and hacker
> skills for the benefit of Debian and the world :-)

Thanks for the lucid overview; after going through it, I think something
close to option 3 would be most effective:

Having separate trees for the yaird core and the debian subdir sounds like
a good idea to me: it underlines the goal that yaird should be written
portable enough to function in other distros.  That portability helps
to find bugs quicker in practise: the ldd problems for instance showed up
in Fedora before they did in Debian.

Letting the debian subdir be maintained by the debian-kernel team is
also OK: interactions between kernel and ramtool are quite frequent.
The amount of changes I make to the debian subdir is small enough to
manage by mailing patches, without having to have write access to the
kernel tree.

As regards the yaird core, I'd prefer maintaining it as a project separate
from debian-kernel, with separate mailing lists.  Motivation: at the moment
I'm subscribed to debian-kernel, and the overload is impressive: it's too
easy to spend all free time catching up on mail, leaving no time to build
new stuff.  Also, it's good to be able to allow people to maintain the
yaird core without having to grant them write access to the kernel SVN.

Where should yaird core be hosted?  It's not pure Debian, but could be
hosted on Alioth under SVN, provided a DD sponsors it.  Hosting the core
on Alioth can only help effective coordination between core and subdir,
so I would prefer Alioth hosting over other sites.

What do you think?

Regards,
Erik




Reply to: