Re: Release team for etch?
El sáb, 11-06-2005 a las 08:15 +0200, Giuseppe Sacco escribió:
> Il giorno ven, 10-06-2005 alle 18:58 -0700, Thomas Bushnell BSG ha
> > Steve Langasek <email@example.com> writes:
> > By the end of June, decide the release criteria.
> > By the end of September, have the GCC changes in place and other
> > infrastructural changes that we know we expect. Have filed bug
> > reports on whatever new RC issues will need to be fixed for etch.
> > (Such as the GCC change, and whatever else.)
> Why do you want a new release without other new important packages like,
> just to make an example, X.org?
> There are a lot of new large packages that needs a *lot* of testing
> before being releaseable. I think a one year release is really
> impossibile, while a two years release seems reasonable to me.
Why do you want a release with X.org packages and not with ..... (fill
in here). The problem is that Debian is quite huge and have packages
with very different schedules and in different cycles of their live. If
we release each 18 months, and for that we have to drop X.org support
from the first release, we will have X.org support in stable in 3 years
form now. If we wait for X.org (or other packages facing a great
transition now), perhaps we will be able to release in 2-3 years, which
will mean that we will only ship one stable version in that period of
time. IMHO that does not benefit users more than having X.org in next
Also, the other problem with large transitions is that they mix with
other packages making transitions too. Packages that won't have a
transition in next year/two years will have it in the middle of the
release schedule... and "why do you want a new release without other new
important packages like, just to make and example, .......?" They will
also need a lot of testing.
That's why large transitions or new packages which will need a lot of
testing need to be treated with care. And X.org have a big advantage...
it can live together in unstable with XFree86
Jose Carlos Garcia Sogo