Re: Bug#15859: libc6 in stable is horribly broken
"Scott K. Ellis" <storm@gate.net> writes:
> On 13 Dec 1997, Martin Mitchell wrote:
>
> > "Scott K. Ellis" <storm@gate.net> writes:
> >
> > > On Fri, 12 Dec 1997, David Engel wrote:
> > >
> > > > On Fri, Dec 12, 1997 at 03:19:29PM -0500, Chris Fearnley wrote:
> > > > > libc6: Conflicts: (libc5<<5.4.33-6)
> > > > > (Necessary due to utmp issue -- Hell, someone upgrading from a CD
> > > > > with stock 1.3.1 will be able to corrupt utmp in the current scheme
> > > > > anyway!)
> > > >
> > > > I can add this in the next release (due very soon) so let me know ASAP.
> > >
> > > Please don't. This will still gratuitously break small upgrades. Adding
> > > a warning about potential corruption should be sufficient.
> >
> > I disagree. The whole integrity of the libc5->libc6 transition will be
> > broken by such hacks, and will keep Debian 2.0 unstable forever if we
> > resort to this.
>
> If libc6 conflicts with every libc5 that can be installed with libc5-dev,
> you've ruined every chance I have of providing a useful workaround to
> people who want libc6 and to keep libc5-dev. I don't consider using
> --force in dpkg a viable alternative.
I don't consider using --force an alternative either. My question is, how
many options do we really want to support? What is the whole point of the
altdev series of packages? To enable libc5 development on a libc6 system.
My primary concern is the stability of systems here. And since this issue
is a choice between either a stable system and a slightly longer upgrade,
or an unstable system and a shorter upgrade, I'll still choose stability.
Martin.
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-request@lists.debian.org .
Trouble? e-mail to templin@bucknell.edu .
Reply to: