I guess the impact will be pretty minimal.
I'd guess _not_! Libc6 installs are always somewhat scary. I'm well and truly fsck'd after installing Mozilla from sid and getting libc6-2.3.1. I can't go back to libc6-2.2.5, because KDM won't start anymore without 2.3, and I can't run anything in Java without 2.2.
Um - my idea was that one first _checks_ what would get updated (by using -s as in simulate) and then would decide if it's worth it. I'm running a
Except you told him that the probable impact of installing libc6 itself,
not its dependencies, would be minimal. This is far from true. I've
_often_ had problems with unstable libc6 installs. It's true I let
KDE/KDM get upgraded, and things would be simpler if I hadn't, but Java is broken purely by installing libc6.
