AFAICT this problem is caused by the change from username `hal' (in
0.5.7-2) to username `haldaemon' (in 0.5.7.1-1).
When upgrading, the 0.5.7-2 postrm script only removes the `hal' user
when purging. The 0.5.7.1-1 postinst script should check for the
existence of a `hal' user and either remove it or rename it to
`haldaemon'.
The warning is being issued because when the postinst script runs
if ! getent passwd haldaemon >/dev/null; then
adduser --disabled-password --quiet --system --home /var/run/hal \
--gecos "Hardware abstraction layer" --group haldaemon
fi
The directory /var/run/hal exists and is owned by another user (hal).
On the system I just tried, even purging hal beforehand does not work
correctly because the dbus packages restart hal when
invoke-rc.d dbus force-reload
is called as part of hal's postrm.
I'm about to try the following fix:
set -ex
if getent passwd hal >/dev/null
then /etc/dbus-1/event.d/20hal stop || true
deluser --quiet --system hal
delgroup --quiet --system hal
rmdir /var/run/hal
fi
--
Trent Buck, Student Errant
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
