Re: debian and linus kernel, the difference??

[Hugo Vanwoerkom <hvw59601@care2.com>]

Brian P. Flaherty wrote:
> Ken Bloom <kabloom@ucdavis.edu> writes:
>
>
> > On Tue, 30 Nov 2004 16:51:24 +0000, Jon Dowland wrote:
> >
> >
> > > On Tue, 30 Nov 2004 10:09:45 -0600, Hugo Vanwoerkom <hvw59601@care2.com>
> > > wrote:
> > >
> > >
> > > > In general I would run WITH the patches, since powers greater than I
> > > > have decided they would be a good idea. Debian certainly is something
> > > > greater than I ;-)
> > >
> > > I'm going to look at what these patches are. Back in the Herb Xu era, I
> > > disliked the volume of backports and somewhat untested stuff that was put
> > > in the debian kernel.
> >
> > Debian Kernel 2.6.8 could burn CD's.
> > Linus' Kernel 2.6.8 couldn't.
>
>
> This seems like a timely discussion because if you check the second
> last issue of kerneltraffic, there is a synopsis of lkml discussion of
> the 2.6 development model:
>
> http://www.kerneltraffic.org/kernel-traffic/kt20041117_284.html#5
>
> Based on the kerneltraffic synopsis, it sounds like kernel developers
> have changed the meaning of 'stable' in the 2.6 series in an effort to
> get things into the kernel faster.  The synopsis suggests that the
> distributions (e.g., Debian) are responsible for making the kernel
> really stable ('really' as in actually stable, not super-stable) and
> kernel developers focus on development. 

Meaning that the patches do that. I posted a synopsis of those earlier:
http://esquipulas.homeunix.com/index.php?p=55

But I would like a simple explanation of them. Note that the "important" 
one in 2.4.x cramfs for initrd is now in the kernel with 2.6.x, as I 
understand it because I don't find the patch.

I use 2.6.9 with Sarge because it has a more or less stable patch for 
Ruby Multi-seat Linux, which fits clean with the Debian patches.

I could use 2.4.27 also but:
That Ruby version disables gpm copy/paste, shows garbage on vc on 
startup with my nVidia video cards, and has no fb support.

On the other hand 2.4.27 has a working SVGATextMode, which fails in 
2.6.x, which I still find much better than fb.

You win some, you lose some.

H















 Furthermore, the
> distributions have all the beta-testers (read as users?) and can
> funnel bug information back to the kernel developers more efficiently.
>
> If this is true, then this may be a good reason to use Debian kernels,
> rather than kernel source from www.kernel.org.