@mvo I don't understand why we'd add the "assumes" to the kernel's snapcraft.yaml now since we're reverting the problematic AppArmor commit. A kernel with the AppArmor commit will never be released to stable (or -updates/-security) so I don't think that the "assumes" workaround is needed any longer.
We decided to revert the commit rather than go down the "assumes" route because the breakage would affect more than just snap confinement. It also affects non-snap processes that are confined by AppArmor.
@mvo I don't understand why we'd add the "assumes" to the kernel's snapcraft.yaml now since we're reverting the problematic AppArmor commit. A kernel with the AppArmor commit will never be released to stable (or -updates/-security) so I don't think that the "assumes" workaround is needed any longer.
We decided to revert the commit rather than go down the "assumes" route because the breakage would affect more than just snap confinement. It also affects non-snap processes that are confined by AppArmor.