AFAIK, catalyst is using sys-fs/squashfs-tools from the hosting FS, not
the one in the chroot.
Seems that there's always the potential of breaking the liveCD, when its
kernel doesn't support the squashfs generated by catalyst.
Would be effective if catalyst performs the assertion that kernel &
squashfs versions agree.
Even if the above is not favorable, I think catalyst should not depend
on the hosting FS for generating an image. I assume a major design-goal
of catalyst is to be as independent as possible from the hosting
Perhaps catalyst can use a chrooted version of squashfs-tools (like it
does w/portage). That would give users a good intervention point for
overriding the version of squashfs-tools they want (by setting it in