Re: James & me @ FOSDEM talking about m68k
On Tue, Feb 27, 2007 at 04:36:11PM +0100, Michael Schmitz wrote:
[...]
> > Later on, we talked about us not being part of etch, and what we could
> > do about that (basically, he asked me whether we were okay with him
> > dropping all m68k packages from stable once it released, which I
> > denied). I mentioned the suggestion that had been done by aj regarding
> > us having an etch and separate britney runs for m68k, but also that not
> > much had happened in that regard; I also talked about the suggestion
>
> *blush* I'm still willing to help there - I won't have much time to figure
> it all out by myself though.
Okay. Same here :)
> > which Jeroen and I had come up with in Breda last summer, where we would
> > still be able to do uploads targeted at an m68k version of stable after
> > the release. James first thought I meant we would like to be able to
> > upload to "regular" stable after it released, which he told me would be
> > impossible without changes to dak to remove a safety net forbidding
> > random changes to stable (which he's not willing to make, for obvious
>
> That has never been an option. What I understood to be possible was
> re-adding an arch once a stable point release is ready.
Right, but that would be in the hands of the stable release managers,
and I guess would depend on how good our port works :)
[...]
> > Personally, I think it'd be best if we could go this way. Having a
> > testing-m68k with our own britney runs and our own criteria for having
> > packages migrate to etch makes it much harder for us to remain in sync
> > with "regular" etch; being able to polish up some packages after the
> > fact makes that much easier. So I'd like to request that as part of
> > releasing etch as stable, ftpmasters create a second suite called either
> > "etch-m68k" which we can upload packages to.
>
> So you want to bypass testing altogether, and instead upload to etch-m68k?
Yes.
The main problem with having our own "testing", as I see it, is syncing
with the testing that already exists, and still being able to do
sourceful uploads that are required for our arch-specific bugfixes. If
we build a stable that's, uh, "etched" (no pun intended) on a stable
release, whereby we upload fixes for m68k-specific bugs using sourceful
uploads (if necessary), I think that'd be way easier to accomplish. No?
--
<Lo-lan-do> Home is where you have to wash the dishes.
-- #debian-devel, Freenode, 2004-09-22
Reply to: