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

Re: More FHS discussion (Was: Uploaded debhelper 2.0.09 (source all) to master)



On Wed, Jul 07, 1999 at 04:22:39PM -0700, Joey Hess wrote:
> Joseph Carter wrote:
> > Um.  Given glibc2.1 it would be Very Unsmart to try to use potato packages
> > on a slink system.
> 
> Arch: all

Yes this could in theory work, but it's a very small subset of packages
and I'd rather it be considered the exception that these things do work
rather than the rule that all things should.

Given that slink features apt and potato actually works best with it, this
will be a moot point.  Just apt-get source what you need and build.  Very
few things would this not work on---those few things including glibc, gcc,
and X, things which are simply not able to be easily built on many
workstations.  Of these, X is the only thing you might really want to
upgrade in that manner.

Everything else can just be rebuild if you need it and it's a hell of a
lot easier to get working that way.


In short partial upgrades of things in unstable should work where possible
(example that I am upgrading potato but perl5.005 is on hold because of
the large number of things depending on unversioned perl) but the same
sort of partial upgrade cannot be expected between stable and unstable
binaries.  This is all we can reasonably be expected to provide.

When slink's epic4 had a DoS in the ANSI color parser the fix was "install
potato's epic4"---but that couldn't be expected to work given potato is
glibc2.1 could it?  In fact it wouldn't, so I rebuilt it on master.


> > I don't know who came up with the idea of partial
> > upgradability
> 
> Partial upgradability has been something debian has always managed. Every
> single person who tracks unstable is using partial upgradability every
> single day.

That's not the same thing as partial upgrades between releases.  A full
upgrade from stable to unstable should work when possible (unstable CAN
break after all) but to the ability to install binaries from potato on a
slink system is insane at best.

I hope this email makes more sense to you than it does to me.  I'm going
horizontal for a number of hours now.

-- 
Joseph Carter <knghtbrd@debian.org>            Debian GNU/Linux developer
PGP: E8D68481E3A8BB77 8EE22996C9445FBE            The Source Comes First!
-------------------------------------------------------------------------
"Now we'll have to kill you."
        -- Linus Torvalds

Attachment: pgp7hQnrdPYqs.pgp
Description: PGP signature


Reply to: