Re: newer kernels from experimental?
On Thu, 2013-03-14 at 03:02 -0500, Mark Allums wrote:
> Some of the 'buntus use 3.4, don't they?
Packages for Quantal are already > 3.4, but it's unstable. However, it
isn't unstable regarding to the kernel. I also build kernels myself for
Ubuntu, it still is buggy as hell. On Arch Linux everything is stable
with 3.7 and 3.6-rt.
In the past I sometimes, very seldom had bad kernels, but most of the
times kernels released by https://www.kernel.org/ were stable. I very
often had unstable versions of distributions installed, but not the
kernel was the reason for the instability. For good reasons I'm not
using Debian at the moment. It's not stable enough for my taste, on my
machine, for my usage and workflow. YMMV.
There always should be a base system, separated from the rest user
space. So if you have a minimal install of what distro ever, this always
should be stable and it usually is.
The short version: If https://www.kernel.org/ says a kernel is stable,
even rt patched kernels, than they most of the times are stable.
Sometimes a new stable rt patched kernel does perform less good, than an
older version, but that doesn't make them unstable.
The advantage of free as in speech *nix OS is, that kernels that are
released as stable usually are stable and that a basic system usually
also is stable.
You can't do anything against cosmic rays from the super nova next door
and you cant provide thousands of user space packages that by any
combination will keep the system stable.
Reply to: