Comment 7 for bug 714356

When I ran my tests, I did notice that it had a little warning about the fact that I was running on batteries, suggesting that it might be safer to plug in.

You suggest removing the power manager. I had considered removing the gnome-power-manager package, but I was under the *mistaken* impression that it would also remove a whole list of other things because of cascading dependencies. I tried again just now, and I see that apt simply removes two packages: gnome-power-manager itself and also "ubuntu-desktop", a meta-package that brings in all of the basic GNOME stuff in a single sweep. This is to be expected, since I am removing one of its dependencies.

So now I have removed the gnome-power-manager completely with no ill effects. Thanks for the suggestion. Now I have two completely suitable work-arounds ("sudo chmod -x /usr/bin/gnome-power-manager" and "sudo apt-get remove gnome-power-manager"). Both of these solutions work for me, because I am on a single-user machine, and so it's OK for me to make my preferences global.

However, this bug report is not about "fixing my system" but about making update-manager work better. So below, I will focus on that.

After removing g-p-m completely, I ran update-manager again. I was surprised that it still gave me the "it's safer to connect the computer to AC power before updating" message. So apparently it is smart enough to figure that out even if gnome-power-manager is not running (or even installed).

So the bottom line for this particular bug is that if you have gnome-power-manager INSTALLED but it's not currently running, for whatever reason, update-manager will run it for you. Note that on a multi-user system, it might be perfectly normal for g-p-m to be installed, but started by some users and not started by others. In fact, now I am curious what would happen if a different user (say, my wife) were logged in using a different desktop environment, something besides GNOME.

So I am not sure if having update-manager start up g-p-m is what one would expect. In fact, the way that update-manager is currently reporting my own battery/AC state even after I removed the g-p-m package suggests that it might not be entirely necessary to query g-p-m at all.

Alan