Re: screen/display switching

[Philip Blundell <pb nexus co uk>]

On Tue, 2003-03-18 at 00:03, Owen Taylor wrote:
>  - A migrated window isn't going to keep it's
>    properties, so "reset _GPE_DISPLAY_CHANGE property
>    to zero length" doesn't make sense to me. Then
>    again, I don't see any reason to need to protect
>    against migrating a newly-migrated window. It
>    should just work.

That was more to cater for the failure case.  If the window turns out
not to be migratable for whatever reason (most likely the target display
was bogus) then it will remain in place with the property set, and it
would need to be cleared down to allow another attempt.  If you take the
view that losing an occasional request by overwriting a nonzero property
isn't a big deal (especially if you are adopting a kind of
fire-and-forget model where the migrated window doesn't report back
status) then this is probably not so important.

>  - Why does authentication need to be in the protocol?

The thinking here is that you might have a shared display, projector or
whatever, which multiple users are connected to.  I don't necessarily
want these guys to be able to steal my windows away, or induce my
application to connect to a malicious server.  Perhaps this is
unnecessary; I'm not sure.

p.