/usr/bin/rootstrap is a python script. It defines three kernel options which
appear to not be used on sid.
'root=/dev/root',
'rootflags=/',
'rootfstype=hostfs',
But then:
# ls -l /dev/root
ls: /dev/root: No such file or directory
Which would imply the script mounts this after it is called. This search
in /usr/lib/rootstrap feels like it's getting warmer:
# grep -n host *
builder:12:os.system("mount -t hostfs hostfs
-o /etc/rootstrap /etc/rootstrap")
builder:16:os.mkdir("/tmp/host")
builder:17:os.system("mount -t hostfs hostfs /tmp/host")
builder:28: configfile = '/tmp/host' +
os.path.expanduser(os.path.join('/tmp/host',os.path.expanduser(argsplit[1])))
builder:30:os.system("mount -t hostfs hostfs -o %s /etc/rootstrap/workdir" %
workdir)
builder:68:os.system("mount -o remount,ro hostfs /")
This would imply a typo in the rootstrap script. But changing root=/dev/root
to /tmp/root changes nothing.
The kernel message implies it can't find init, yet init is passed on the
command line. In order to execute it, the UML kernel would have to aware
of /usr/bin/python. My diagnosis is it isn't.
Other notable discoveries are that Matt Zimmerman has many bugs approaching or
over a year old, yet in /usr/share/doc/rootstrap/changelog.Debian.gz says
he's updated the package in 2004 and mentions Sarge.
--
-l[e^2]
-----------------
chmod -R 777 /
chown -R nobody /
imagine true freedom
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
