New 2.6.0 naming

[Niki Guldbrand]

On Thu, 18 Dec 2003, Moritz Heiber wrote:

> On Thu, 18 Dec 2003 21:34:49 +0200
> Hendrik Visage <hvisage at envisage.co.za> wrote:
>
> Next time please write your answer to the lunar-dev instead to the
> lunar-user list.
>

Moe, Hendrik isn't on our devel ml... So i have cc him here... I belive
that it was elaine that sent it to lunar at l-l.org...

> > On Thu, Dec 18, 2003 at 09:31:56AM -0500, elaine wrote:
> > > Keep the name, it's just a name, a rose by any other name
> > > still has thorns :-)
> > >
> > Whoa, you *are* pessimistic... but not as pessimistic/paranoid....
>
> But absolutely true.
>
> > > On kernels, I'd like to see some experimentation with snapshots
> > > of glibc+nptl. NPTL was merged into glibc back in september and
> > > is probably getting close to release.
> >
> > Which will be adding two (three??) unknowns into the picture,
> > ie. (1) kernel (2) fancy glibc (3) driver integration/porting (ala
> > nVidia etc.)
>
> Yes, since we want a full featured, stable and fast environment.
>
> > > I think the right time to release a new 'stable' 2.6 kernel into
> > > moonbase is after we have a working & stable glibc+NPTL.
> >

I can only agree on this one, and I installed our 2.6.0-test11 on a fresh
installed system the other day, and it went without issues, and i' running
it as we speak... so the biggest issue is proberbly the apps that don't
compile against 2.6, and gettin glibc to runcorrectly.

And i'm writing all this on a machine without a blinking cursor, so it's
realy hard to see where your cursor is while you write ;-)


> > Make some sense, however, rather get people to get a stablish 2.6.x
> > working sooner rather than later, as (3) above is going to be fun, and
> > then you have the NPTL to content with too.... nVidia's drivers
> > already have fun with the nptl (afair)  and non-nptl versions....
> >
> > So rather try and get the kernel at least running as best as possible,
> > then add the rest.
>
> Nope, I disagree.
>
> a) We'll have to wait for Nvidia to release a new driver that is
> rather running out of box in connection with 2.6.x than using ugly third
> party patches (Yes, I know .. they're working ... but still).
>
> b) Having a good glibc integration is half of the way for wining the
> price. I don't want to force two glibc recompiles for every user out
> there just because we had to commit unfinished business.

her her

> c) Read the article about 2.6 at http://www.kerneltrap.org (Especially
> Andrew's "MUST FIX" and "HAS TO BE FIXED" lists). You'll get the idea.
>
> Moreover a close friends of mine is doing support at Gentoo's forums and
> I know that he and his fellow friends are getting absolutely nuts about
> 2.6 kernel users complaining about issues that are either kernel related
> (Bugs that are yet to be solved) or incompatibilities with the new
> device/information structure of the new kernel.
>
> So the list is:
>
> 1) Test the new kernel, even be on the lookout for 2.6.1 (I expect it
> for mid January).
> 2) Integrate and test the new glibc.
> 3) Integrate and test the new (yet to come) Nvidia drivers.
> 4) Xfree? $ISSUE_WITH_2.6
> 5) Push it out.

I belive that the upcomming release of Xfree is 2.6 safe (that is what i
have heard...). But else all this sounds resonable...

> Lets avoid all that "2.6 hype" and do some real testing on this new baby
> before getting it out.

Her her... I'l keep running 2.4.xx on my systems until i'm confident that
2.6.xx is good enough, and i'll keep the work on them for the time being,
i know that csm has asked to get 2.6 support asap, and i'm willing to do
test on my devel ssytems with it, and i any one (devel) wants access to
some of them i'll gladly provide that...

Don't know if i have more to add at pressent...Other that it's realy
anoying to have to give up your devel machine... :-( I hope i get it back
next week befor chrismas ;-)


Niki

> Bye,
> Moritz
>
> --
> Moritz Heiber
>
> Lunar Linux developer: http://www.lunar-linux.org
> XFce4 desktop environment: http://xfce.org
> Moritz Heiber[at]SPLiNe: http://moe.spline.de
>