Re: trimming packages
Op za, 16-04-2005 te 17:40 -0700, schreef Rick Younie:
> I think that we should move toward only building packages
> that have some use for m68k, not just building packages to
> show we can do it (eventually.)
No, I disagree. First, it's very hard to tell what exactly 'some use'
is; we can't honestly make that decision without making a judgement call
that, in the end, should be made by our users rather than by us. Second,
many of those packages that have 'no use' have an entangled,
intertwined, and ever-changing build-dependency maze between them and
other packages that most certainly /are/ of use to m68k users;
therefore, I think going down that road will likely make our life much
harder, rather than easier.
That being said, if we should only do what is "useful", we might just
end here and now. Nobody in their right mind is using an m68k-based
system as their main box these days, so an m68k-oriented distribution
isn't really "useful".
> And I don't think adding
> more 060s is going to help us. We do have a place in the
> new structure but we'll die a slow death if we try to
> continue as we've been doing.
If that happens, we'll have plenty of time to change our strategy. I
don't think we should do that because someone proposed some change.
As I understand now from discussion on -devel, the changes aren't as
intrusive as I first thought. We'll most likely be moved to a separate
mirror network (which is reasonable), and if we comply with some quality
requirements, we'll be able to do full releases (which would be
reasonable if the quality requirements were, which they aren't yet; but
I'm sure we can get them to be reasonable given some more discussion).
In addition, Martin Schulze (of the Security Team) stated that for the
team, given working security autobuilders, it doesn't matter whether
they need to support 2 or 20 architectures--so we will get security
updates, too.
Considering all that, I'd say the proposed scheme is much more
reasonable than I first thought, and that we should support it (given
some, IMHO, minor changes); but I don't think a change in our build
policy is called for.
--
The amount of time between slipping on the peel and landing on the
pavement is precisely one bananosecond
Reply to: