Re: testing and no release schedule
On Fri, Mar 26, 2004 at 02:31:20PM +0100, Adrian Bunk wrote:
> > How about we just stop making excuses? If dead architectures are holding up
> > the release process, then drop them, or make them secondary platforms. m68k
> > wouldn't be any worse off today if we had released an i386 sarge six months
> > ago.
>
> Which of the architectures Debian 3.0 released with do you consider to
> be dead?
A lot of them. However, that's not the point. The point was to stop making
excuses. If niche (politically correct term for dead) architectures are
holding up the release, then let's release without them.
> A real problem are RC issues in essential packages like e.g. glibc.
>
> Let's check for which architectures there are architecture-specific RC
> bugs open against glibc today [1]:
> - i386
> - alpha
> - sparc
> And there are also benefits of supporting many architectures, an
> example:
>
> hppa in Debian 3.0 uses gcc 3.0. Therefore, the hppa maintainers had to
> report all compile errors with with g++ 3 as non-RC bugs.
>
> Some bugs were not fixed before Debian 3.0 was released, and this was
> only a hppa problem. But the big amount of reported and mostly fixed
> issued with g++ 3 made the g++ 3.2 transition for all architectures much
> easier.
I think we are in agreement that this is a management problem, and not
necessarily a manpower problem. I also don't know if it's a solvable
problem, since solving it will require a change in Debian's culture at the
highest levels.
--Adam
--
Adam McKenna <adam@debian.org> <adam@flounder.net>
Reply to: