Dave Jones wrote:
On Wed, Jan 18, 2006 at 06:23:48AM +0800, John Summerfied wrote:
> Fedora Project isn't providing as good a service as it could, with just
> a teensy bit more effort.
I think you underestimate the effort involved with maintaining another branch
of packages. Bumping up to the newest version of a package isn't always
feasible, and it's a ton of work backporting fixes to older releases.
Tracking three streams of development is a nightmare (And this doesn't
even count RHEL streams that your average Red Hat developer works on
in parallel).
Handing off releases to legacy has been a great way of freeing up developer
time so that the stability of the current release increases and
the next release gets into a state ready for shipping.
Just the amount of time saved doing bugzilla triage on FC3 is going to pay
off huge now that it's reached EOL. I got the kernel bugcount pretty low at EOL,
and now I can go about doing the same thing for FC4, which has been suffering
at the expense of FC3 for a while.
Dave
Dave,
we're only talking about a short period of support. If Rahul is right
(and I don't believe he is), there won't be any fixes.
If he's wrong, I don't imagine that there will be many fixes, so we're
talking about only a little bit more support - a short period of time,
maybe a few important fixes.
And those fixes are being generated anyway: all that's needed is better
timing and better coordination. I really don't see much change to the
workload at all, and I can imagine it might even be better: do the
handover when the panic of pushing the new release out the door is over
and things have calmed down a bit, instead of when the excitement's
building and everyone's running round trying to get last-minute changes
in, trying to justify rebuilding everything so we can get the latest
prerelease gcc or glibc or whatever in.
--
Cheers
John
-- spambait
1aaaaaaa@xxxxxxxxxxxxxxxxxxxxxxx Z1aaaaaaa@xxxxxxxxxxxxxxxxxxxxxxx
Tourist pics http://portgeographe.environmentaldisasters.cds.merseine.nu/
do not reply off-list
