Re: [PROPOSAL] Debian Release Plan
On Saturday 02 August 2003 09:01, Alastair McKinstry wrote:
> I disagree. We should ship ASAP despite, or even because of, older
> milestones. With RC bugs and d-i (as is) fixed, Sarge would still be an
> improvement on current stable, woody: the longer between releases the
> less useful the distro is, as it lacks modern drivers on the CDs.
> Already people are running into problems installing woody due to old
> drivers: eg new servers with gigabit NICs not supported in woody CDs
> make installing very painful.
But the very same applies to XFree86 and supported graphics boards.
> Secondly, we need to signal to upstream to fix up _their_ act, too. If
> we can't ship, for example the latest gcc because glibc isn't ISO C
> compliant and working with gcc-3.3 (see other thread), then others need
> to act: glibc maintainers (upstream). Why is it considered OK for other
> commercial distributions to ship shoddy software? Instead of being
> ashamed of shipping old versions, we should ship whats in testing, and
> let people ask questions as to why we're not shipping gcc 3.3. And
> answer them.
Yes. One implicit part of the development is however that many projects
advance much faster than older maintained versions, not only in terms of
features, but also in terms of bug fixes. For lots of projects the saying
"it's old but proven to be rock-stable" does thus not apply.
I'm not talking about the one-liners (which can be applied to stable branches
even in upstream CVS), but about redesign and/or rewrites. There's probably a
difference between 'stable' and 'mature software'.
Apart from that, packages which don't go into testing because of compilation
errors shows IMO that whenever there's an upgrade of the default compiler,
some weeks are always needed for problems to settle down.
Play for fun, win for freedom.
Hurd^H^H^H^HLinux-Info-Tag Dresden 2003: http://www.linux-dresden.de