Re: irc meeting regarding kernel status d-i RC3
On Sat, 05 Feb 2005, Joey Hess wrote:
> Andres Salomon wrote:
> > I'm of the opinion that if sarge is literally less than a month away from
> > freezing, then let's stick w/ 2.6.8. It's Good Enough; if we can get
> > around ACPI problems by disabling it, fine. However, if sarge is still
> > far away from freezing, then there's no reason not to go w/ 2.6.10.
> > It's clearly shown itself to be a better kernel in some cases; we just
> > need testing to find regressions in *core* stuff (individual drivers,
> > filesystems, etc, can all have fixes backported. It's the core stuff
> > that's difficult). I'm not concerned about when d-i rc3 is released; I
> > care about when *sarge* releases. People said that once rc2 was out, sarge
> > would be ready. That's clearly not the case, and probably won't be the
> > case once rc3 is out.
> b) Many of the people whose time would be spent working on updating d-i
> to a new kernel could better use that time to work on other RC issues
> for sarge. For example, I work on tracking security issues in testing.
> Kamion and vorlon work on getting the release out. Etc.
i agree with dilinger:
if the timeframe is reasonable soon, let's stay with 2.6.8.
but 2.6.8 has many problems appart the one discussed,
a good example is #292478.
backport of the usb system is out of question due to
potential destabilisation + core changes.
2.6.10 needs to be considered if time frame shifts.
1/2 a year of kernel dev resulting in better hardware support
(acpi newer intel board) should not be thrown away too easily.
hope that sparc and other arch catch up, so that 2.6.10 remains
an option. fc3 kernel is based on it. it had a longer stabilisation
phase than previous 2.6
the other bonus is the -as tree from dillinger, which is used
by other distribution and collects many nice fixes for 2.6.10.