Re: bonobo activation environment matching and DISPLAY

[Mark McLoughlin <markmc redhat com>]

On Tue, 2004-08-24 at 11:54, Michael Meeks wrote:
> On Mon, 2004-08-23 at 17:22 +0100, Gustavo J. A. M. Carneiro wrote:
> > > I assume that your note above about the bonobo:environment property
> > > doesn't mean that the "artificial environment", i.e. the constructed
> > > req_env, is deprecated...  let me know if I've missed something.  
> > 
> >   Yes, I think it is deprecated, but I defer to Michael on this one. 
> > Michael, what's your opinion on this?
> 
> 	I believe the new idea was to have each .server file denote which bits
> of the client environment it really cared about; and thus have the
> server process determining it's own per-display/per-whatever lifecyle
> rather than the calling-client. That would seem to be a good idea in
> general - I forget the precise syntatic details, but it focused around a
> bonobo:environment stringv in the .server file AFAIR.

	I think there are few problems here:

 - The $DISPLAY in the environment isn't canonicalised so comparing 
   semantically identical $DISPLAY in different process might not match.

 - The $DISPLAY in the environment doesn't necessarily reflect the 
   actual display being used - think about --display and --screen.

 - Merely keying of the $DISPLAY in the environment is not necessarily 
   always what you want - e.g. if you want per-screen instead of 
   per-display or vice-versa

	So, the old API wasn't perfect - but there were reasons for it. I think
the main thing we were missing was an easy way of setting DISPLAY with a
canonicalised value without or without the screen number.

Cheers,
Mark.