Re: [D-I] Supporting 2.6.14 kernels in base-installer
On Mon, Nov 14, 2005 at 03:47:57PM -0500, Joey Hess wrote:
> Sven Luther wrote:
> > Ok, but who will chose which will be the default one ? The d-i team, or the
> > kernel team ?
>
> It's more important to make a choice than what the choice is.
>
> d-i chose between lilo and grub before; we did it by listening to what
> users wanted and looking at which worked best overall, and I think we'll
> eventually get the same kind of information about initramfs-tools vs
> yaird (the wiki page is a good start). Before we decided to default to
> grub, we temporarily, and somewhat randomly, defaulted to lilo for a
> long time.
>
> This seems more in-scope for the kernel team to decide than the d-i
> team, but if it would help I can flip a coin..
Indeed, but the kernel team seem to agree that none of the tools are really
mature enough at this time, and that we want to make the final choice later,
when we have more background and experience, and the tools have gotten more
maturity.
What does that mean for d-i ? I don't know exactly, but i will look at Colin's
change with interest, since he plans to add initramfs-tools support.
> > And what about the cases where the choice happens to be dependent on
> > the architecture used ?
>
> As far as I can tell so far they both have the same set of supported
> architectures, although initramfs-tools has been tested as failing on
> more architectures. But if we really needed to, we could vary it by
> architecture without most of the complications that offering a choise at
> install time would entail.
>
> (BTW, should we let the new kernel packages that require these
> tools into testing, if they don't yet support most architectures?)
That is why an RC bug is open against the 2.6.14 kernels, so that this will
not happen until we all judge it ready. What does this mean for d-i and the
way it handles etch/sid separation i don't know. I feel it would be best to
start building sid-d-i images installing sid for a month or two to get more
feedback about those issues.
Also, about choice, my though was to enable both in d-i, and be able later
easily enough to go one way or the other.
Friendly,
Sven Luther
Reply to: