On 09/12/2012 05:30 PM, Roger Leigh wrote: > We should probably wrap all of the restart logic (except the inittab > stuff) in ischroot checks. Messing with the host's /dev is not > worth risking. There was a reason why we did it this way though; it > wasn't unintentional. I'll have to remember the reasoning, since I > don't want to break something else by accident. Possibly vservers/lxc > or some other chroot-like thing.
yep, i figured that was the case, but i don't remember the reasoning any
more than you do :) perhaps digging through some revision control logs?
> However, why is the host's /dev accessible when bootstrapping--it
> certainly shouldn't be bind mounted. Is this something which
> fakechroot is doing?
debootstrap's fakechroot variant uses a symlink to pseudo-bind-mount the
host's /dev/, iiuc.
hth,
--dkg
signature.asc
Description: OpenPGP digital signature
