Re: [rox-devel] Re: ROX in Pager/Tasklist -- Problem (maybe) analysed

[Tony H. <h...@re...>]

In <200...@ev...>, Thomas Leonard wrote:

> On Sun, Jun 13, 2004 at 07:00:48PM +0200, "Daniel Pörsch" wrote:
> > 
> > But grepping on the source I found
> > void make_panel_window(GtkWidget *widget) {
> > // ...
> > 		state_list[0] = gdk_atom_intern("_NET_WM_STATE_STICKY", FALSE);
> > 		state_list[1] = gdk_atom_intern("_NET_WM_STATE_SKIP_PAGER",
> > 						FALSE);		state_list[2] = gdk_atom_intern("_NET_WM_STATE_SKIP_TASKBAR",
> > 						FALSE);
> > // ... and later
> > 	gdk_property_change(window, xa_net_state, xa_atom, 32,
> > 			GDK_PROP_MODE_APPEND, (guchar *) state_list, 3);
> > 
> > which should as far as I understand set _this_ settings. So I think the
> > question we should ask is, what happens with this settings?
> > 
> > Unfortunately I am no C-programmer and don't know how to debug this, but I
> > hope this help resolving the bug.
> 
> What version of GTK are you using? Probably, some combination of GTK and
> the window manager are messing with the properties...

I've got that problem with libgtk-2.4.1-4 and both OroboROX and
metacity. At one time, not having spotted the above code, I added calls
to gdk_set_skip_{taskbar|pager}_hint, somewhere in panel.c IIRC, which
cured it. 

A little further on in make_panel_window() there's this comment:

    /* Note: DON'T do gtk_window_stick(). Setting the state via
     * gdk will override our other atoms (pager/taskbar).
     */
                                                                                
Maybe some clue there? How about trying applying the _NET_WM hints after
the wm_hints_values?

-- 
TH * http://www.realh.co.uk