Re: 2.6.13, experimental and 2.6.14-rc ...
On Tue, Oct 04, 2005 at 09:19:58AM +0200, Sven Luther wrote:
> On Mon, Oct 03, 2005 at 06:01:53PM +0900, Horms wrote:
> > On Mon, Oct 03, 2005 at 06:18:40AM +0200, Sven Luther wrote:
> > > Hi,
> > >
> > > I have just been thinking about something this morning.
> > >
> > > We apparently need to wait for 2.6.13 to be uploaded for some time yet for the
> > > debian-installer beta release to happen. So, i wondered if we would not be
> > > better off by forgetting about 2.6.13, and make a 2.6.14-rc upload to
> > > experimental asap, and work on that for a while, while keeping 2.6.12 in
> > > etch/unstable ? This would allow to perfection the initramfs-tools/yaird
> > > migration and then come up with a 2.6.14 upload the day of the upstream
> > > release hopefully.
> > >
> > > Anyway, i will be away and probably netless the next week, so i just wanted to
> > > throw out the idea about this before i leave.
> > I'm cool with using experimental if we have to stick with 2.6.12
> > in etch/sid in the short term. But do we have any idea how long this
> > is going to be. Are we talking in terms of days, weeks or months?
> > In the case of months, I guess that means we should be looking at
> > a fresh 2.6.12, or does that have a negative impact on the
> > packages in testing situation?
> Well, the constraint is with the debian-installer folk on this, but i get the
> impression that it is at most a week or so, but better ask Franz Pop about it.
I've CCed debian-boot, hopefully Franz or another d-i person can comment
on this. I'd like to get a better understanding of what the issue
relating to 2.6.12 in testing and d-i is. Though I do not see a week or
so as being a major issue.
> > As for 2.6.13 vs 2.6.14-rc, thats really a bit of a moot point.
> > My feeling is that by they time we get 2.6.13 ready, 2.6.14 will
> > probably be out, so aiming at 2.6.14-rc for now might be a good idea.
> > But I'm really more concerned about making sure that the > 2.6.12
> > release happens. Walidi and Dilinger were doing some fine work in that
> > area, but it seems they are busy and not much is happening.
> Ok, that seems to correspond to my own feeling, so let's try to do a 2.6.13
> upload ASAP to experimental. I am also busy this week, but we should go ahead
> on that next WE.
I'm happy to go with experimental any time. Though time is something
I have none of right now :(
> > What is the status of replacing initrd. Do we actually have a solution
> > that works across a wide range of architectres? Wide enough for
> > an upload to experimental? Wide enough for unstable and testing?
> An upload to experimental will allow to sort this out as we go. We need to
> test it, and people can sort out yaird/initramfs on their own arch with
> known-to-work packages in experimental.
> I would prefer to have real feedback for the unstable/testing upload though,
> which is something we will only get by making the experimental upload.
> Alternatively we can just upload 2.6.13 as-is to experimental now, and fix
> things as they go, and migrate it to 2.6.14-rc once people get time, but it is
> important to get packages out to experimental.
My impresion of experimental, is that it is very green, and as such
uploading might not be a good idea. It could help to isolate problems.
Though does anyone have any time to deal with the bug reports in
a meaningful way? If not, then there might not be a whole log of point
to the process.
As for unstable/testing, that is defianately going to give us
a bigger audience. We just need to find time to get the packages
together and time to real with user's responses.