I filed a bug against gnome-power-manager a little while
ago because I could not suspend. It turned out my user was
not in the powerdev group.
It was Joss' initial belief that the non-root user that you
create in d-i is added to a fixed set of groups, including
powerdev. For the remainder of this mail, I will refer to
this user as the "primary user", for clarity's sake.
I did several subsequent installs and my user never ended
up in powerdev (nor netdev for that matter). It's my belief
(yet to check d-i code to confirm) that the user gets added
to powerdev if you select the desktop task: for each of my
installs I only chose standard system and then subsequently
installed a desktop environment (a subset of the task).
I think it would be nice to explore better ways of handling
group membership for users, in particular the "primary
user", or perhaps primary users plural, in time for squeeze.
If there was a system-wide way of defining a set of primary
users, packages could then make the decision to add those
users to groups at install time. So pm-utils could add the
primary users to the powerdev group; fuse-utils to the fuse
group, etc. Perhaps another group could be used to define
this set of users: the relevant postinst would inspect the
group "primary" and add any users in that group to the
This is me thinking out loud rather than proposing a
well-rounded solution. I'm curious who else thinks that
there is room for improvement here? Can anyone see some
immediate flaws with my back-of-the-envelope idea above?
Finally can anyone with a deeper insight of the issues
explain whether or not the frustrating "existing logins
don't inherit new groups" behaviour is fixable, or is that
deeply rooted in UNIX tradition? (I note that it seems
HAL makes an on-invocation group check for suspend so
adding a user to the powerdev group and attempting a
suspend from a pre-logged in session works)
 I don't think this is a great name, suggestions welcome