On Tue, May 02, 2006 at 10:10:29AM -0600, Archaic wrote:
> On Tue, May 02, 2006 at 11:03:42AM -0500, Bruce Dubbs wrote:
> >
> > At some point, we need to freeze the kernel, just like every other
> > package. Users are free to upgrade to the latest of any package of
> > course, but we need a constant platform for testing. The best way to
> > test, IMO, is via building and using the packages in BLFS. That takes
> > some time.
>
> Agreed on all points, especially testing and freezing kernel version.
> ICA proves reproducability, not sanity. LFS rebuildability is a step
> higher in proving sanity and to some degree stability. And an extensive
> BLFS installation can establish sanity and stability. We would need huge
> compile farms to try and reach any level of confidence if the base
> system kept changing.
>
If I'm reading you correctly, that is a pity. 2.6.16.13 is out now
with one fix:
NETFILTER: SCTP conntrack: fix infinite loop (CVE-2006-1527)
I still say that while the -stable team is updating the kernel
version we have in the book, it should be a no-brainer to use their
fixes.
Ken
--
das eine Mal als Tragödie, das andere Mal als Farce
--
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page
