SOLVED! Pidgin ignores DISPLAY variable.

[Rogier Wolff]

On Thu, Jan 22, 2009 at 06:48:10PM -0600, Kevin Stange wrote:
> Rogier Wolff wrote:
> <snip>
> > 
> > Pidgin however doesn't seem to support the first method (no -display
> > option), and it more or less ignores the DISPLAY environment variable:
> > 
> > it mumbles: 
> > 
> > Xlib:  extension "RANDR" missing on display ":1.0".
> > 
> > and then displays on :0.0
> 
> Pidgin has --display= and honors the DISPLAY environmental variable.
> 
> When I try to use an alternate display Pidgin seems to work fine with
> it.  I know it handles multiple displays properly as I've used it on
> multi-user systems with multiple people running Pidgin and over X
> forwarding.  I was able to successfully use it with Xnest.

I was actually trying to use it with Xvfb. 

% Xvfb :1 &
% setenv DISPLAY :1.0
% xterm &
% sawfish &
% pidgin --display :1.0 &
% pidgin &


I have a feeling pidgin contacts some sort of gnome-session manager,
which tells it: "we are SUCH a happy familiy here on :0.0 come join
us!!"

I just tested with Xnest. Same behaviour: 

% Xnest :2 &
% setenv DISPLAY :2.0
% xterm &
% sawfish &
% pidgin &
% pidgin --display :2.0 &

OK Found it. 

It is not connecting to some gnome-thingy, it's contacting another
pidgin. I apparently closed the pidgin window, but it was still
running in the background.

killall pidgin fixes things. Sorry for the confusion. 

May I make a suggestion then? If I'm starting a pidign that determines
anohter pidgin is running with a different display, it should print a
warning....

(Firefox is also annoying like this. If I log into a different machine
because I want to test something on a different firefox version, it
contacts my local firefox, and launches a new window there.... Fine,
if I just wanted to have a new window, but annoying if I wanted to
test something on the slightly different firefox version.)

	Roger. 



> Pidgin would report something like this for an invalid display because
> it uses GDK to deal with X:
> 
> (pidgin:6735): Gdk-CRITICAL **: gdk_display_get_name: assertion
> `GDK_IS_DISPLAY (display)' failed
> ** (pidgin:6735): WARNING **: cannot open display: unset
> 
> So I feel as though your Xlib or GDK is doing something funky.
> 
> > 
> > As far as I know you need to go pretty extensive lengths before you
> > can get an X application to ignore the DISPLAY varibable. Who decided
> > that pidgin should always use :0.0 ?
> 
> This is not something anyone decided to do on our side as far as I know.
> 
> > 
> > I just googled for what RANDR is... the RandR extension is not
> > essential to pidgin. Maybe pidgin likes to be notified of resolution
> > changes, Fine. If not available, that should NOT be fatal. And even if
> > it is, then it should not ignore my DISPLAY setting and display
> > somewhere else.
> 
> I don't believe we use the RANDR extension at all.  GDK tells us about
> screen geometry when we need it, which is rare.
> 
> Kevin
> 
> 



-- 
** R.E.Wolff at BitWizard.nl ** http://www.BitWizard.nl/ ** +31-15-2600998 **
**    Delftechpark 26 2628 XH  Delft, The Netherlands. KVK: 27239233    **
*-- BitWizard writes Linux device drivers for any device you may have! --*
Q: It doesn't work. A: Look buddy, doesn't work is an ambiguous statement. 
Does it sit on the couch all day? Is it unemployed? Please be specific! 
Define 'it' and what it isn't doing. --------- Adapted from lxrbot FAQ