Getting ready for PolicyKit 0.9

[Michael Biebl]

2008/6/30 David Zeuthen <david at fubar.dk>:
>
> Hey Michael (and mailing list),
>
> So I want to release PolicyKit 0.9 this week (not a lot of new features,
> mainly bug fixes). A while back we had a private exchange where you
> pointed out some minor (not security critical or anything) bugs in the
> modes/permissions and the output text from configure. You also asked for
> the rationale for these modes/permissions.
>
> I've finally corrected this with this commit:
>
> http://gitweb.freedesktop.org/?p=PolicyKit.git;a=commitdiff;h=785a63bcb175bc9eee14c664353e82c5ca8a6206
>
> Does this look OK to you?

Hi David,

thanks for looking into this.
I checked the configure output, the installed files/directories and
the README, and they all seem to be consistent now. The explanations
in README are concise and well understandable. Nice work!


On a completely different topic though:
The error message, when CK doesn't return an active session, is suboptimal.
Example: nm-connection-editor from NetworkManager 0.7 uses
PolicyKit(-gnome) to modify system wide settings. If I kill the
running console-kit-daemon instance (so ck-list-session returns
nothing for the currently active user), I get the following error
message (on the console) within nm-connection-editor:

[WARN 21823] polkit-action.c:211:polkit_action_set_action_id():
polkit_action_validate_id (action_id)
 Not built with -rdynamic so unable to print a backtrace

A normal user will not understand what the underlying problem is.

Imho it would be much better, if polkit-gnome-manager would show an
understandable error dialog, that the current user is not listed as
being active, with the adive to check the ConsoleKit installation and
logout/login again.

Cheers,
Michael

-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?