Re: Method "getActiveDevice" doesn't exist



On Mon, 23 May 2005, Bastien Nocera wrote:

> On Fri, 2005-05-20 at 22:26 -0400, Dan Williams wrote:
> > On Fri, 20 May 2005, Bastien Nocera wrote:
> > > On Fri, 2005-05-20 at 16:10 -0400, Dan Williams wrote:
> > > > On Fri, 2005-05-20 at 20:32 +0100, Bastien Nocera wrote:
> > > > > Hey Dan,
> > > <snip>
> > > > > How can I debug the notification icon, to know why it's saying I don't
> > > > > have any devices?
> > > > 
> > > > First of all, you're sure you're running 'nm-applet' ?
> > > 
> > > Yup. I tried to though, and it just won't lock the vpn icon, or the
> > > progress ones (patch for better error reporting below).
> > > 
> > > So, how do I debug the notification area stuff then?
> > > 
> > > > Second, the nmtest tool is actually wrong, I never fixed it up after the
> > > > removal of getActiveDevice.  I'll have to do that.
> > > 
> > > Right, but still, it lists the devices and networks, and the
> > > NetworkManagerInfo doesn't.
> > 
> > Put an fprintf() before the dbus_pending_call_set_notify() call in 
> > nmwa_dbus_update_nm_state() in applet-dbus-devices.c.  That will tell you if the 
> > applet is even talking to NetworkManager.  Are you sure you restarted the 
> > messagebus daemon after installing the new copy of NetworkManager from CVS?  If 
> > not, dbus won't be aware of the applet's security policy and the applet will 
> > simply get kicked off the bus.
> 
> Was my stupidity, I should have ran "gtk-update-icon-cache
> -f /usr/share/icons/hicolor/", instead of the command on its own (which
> I thought had the same effect, but I guess some clock screwage threw the
> code off balance).

I had problems with this for 2 days, I couldn't figure out what the hell it was 
supposed to actually do or how it did it.  gtk-update-icon-cache isn't exactly 
transparent.  Anyway, it turns out the measure of "invalid" cache is the mtime 
of the directory.  Because when you install an RPM the mtime changes, RPM 
installs that include gtk-update-icon-cache in the %post usually don't have to 
care about -f.

Dan



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]