Re: Support for WM spec

[Owen Taylor <otaylor redhat com>]

Alexander Larsson <alla lysator liu se> writes:

> On 27 Feb 2001, Havoc Pennington wrote:
> 
> >
> > Havoc Pennington <hp redhat com> writes:
> > > > +#ifdef NO_GDK_WMSPEC_SUPPORTED_YET
> > > > +  if (!_gdk_wmspec_supported (gdk_atom_intern ("_NET_WM_ICON", FALSE)))
> > > > +    return FALSE;
> > > > +#endif
> > >
> > > Possibly should set the icon anyway, in case a supported WM appears
> > > later, as Tim pointed out for the stuff I was doing.
> > >
> >
> > Thinking about it more, we can probably fall back to setting the icon
> > to a GdkWindow that renders the pixbuf, the way you did it
> > pre-WM-spec.
> 
> Yes, having both _NET_WM_ICONs and GdkWindow icon may be confusing for the
> WM.

The WM spec needs to be clear about what setting both _NET_WM_ICONS
and an Icon pixmap or window means.

If _NET_WM_ICONS and a pixmap are set, it's pretty clear that 
_NET_WM_ICONS is the preferred information.

But it's conceivable that if _NET_WM_ICONS and a icon window are
set, then the icon is an active icon, and should be preferred
over _NET_WM_ICONS, and the app is just setting _NET_WM_ICONS
for the information of tasklists and suchlike.

Does there need to be some disambiguation between these, or 
can we just pick one interpretation and ignore the other?

What do window managers do currently?

Regards,
                                        Owen