Torvalds:requiring root password for mundane things is moronic
awilliam at redhat.com
Mon Mar 5 22:00:40 UTC 2012
On Sat, 2012-03-03 at 14:07 -0500, David Zeuthen wrote:
> > On Fri, 2012-03-02 at 08:42 -0600, Greg Swift wrote:
> > > I experience a similar scenario. On my home system (f16) I have my
> > > wife and both in the wheel group. Every time I go to run
> > > virt-manager
> > > I get prompted for her password. I do believe she is first in the
> > > wheel group after root in /etc/group. However this doesn't make
> > > any
> > > sense to me. It makes more sense for users that need that level of
> > > access to all know the root password rather than the users to know
> > > another user's password. Even then, if I am in the same group,
> > > doesn't it make more since to either prompt for my own password or
> > > just allow me? We know each others password so i've always
> > > shrugged
> > > it off cause I'm looking at other issues the few times when I am
> > > playing with the virtuals at home but since someone brought it
> > > up...
> > This sounds pretty straightforwardly like a bug probably in
> > PolicyKit,
> > to me. It's obviously more correct to use the current user's
> > authorization if it's sufficient than just to go with the first user
> > in
> > the admin group in all cases...
> > So, file a bug against PolicyKit.
> (Ugh, no, please don't tell people to file bugs against polkit
> unless you are actually sure it's a polkit problem. In this case
> it's not.)
Sorry about that, but my general take is that it's important to get
issues filed, and it takes about fifteen seconds for a developer or
appropriately clued-up triager to re-assign a bug or mark it as a dupe,
if they know where it should go. So I tend to err on the side of getting
things filed against a product that's *approximately* correct - as in
the person who owns it will at least know where it should go to, if it's
wrong - rather than worrying so much about getting the assignment
precisely correct that the bug never gets filed.
> for details. If the problem is that both users are in wheel but you
> are asked to authenticate as the user who is not logged in, well,
> that's solved in a gnome-shell update, see
That's the bug being discussed here, AIUI.
> and check if that patch is included in whatever version you are using.
Thanks for the reference. I found it independently after my mail, and it
seems the patch should be in F17 but not F16. I asked in the bug if it's
too disruptive to be backported to the stable Shell branch that F16 is
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
More information about the devel