User bus conclusion

[Colin Walters]

On Tue, Nov 9, 2010 at 6:48 PM, Lennart Poettering <mzqohf at 0pointer.de> wrote:
> On Tue, 09.11.10 23:23, Matthew Johnson (dbus at matthew.ath.cx) wrote:
>
>> On Tue Nov 09 17:06, Ryan Lortie wrote:
>> > ...
>> > All of this means that there is only one graphical login per user.
>> >
>>
>> Sorry, I'm I misreading this? Or have you just suggested that we stop
>> supporting one user having multiple logins?
>
> Multiple logins will continue to be supported. But only one of them will
> be graphical.

I think the position is simply that it's up to the environment to
support this, if desired.  For example, a host environment could
detect when it's launched under a second X server, and run multiple
processes (or multiplex X connections), coordinating this through the
now user-machine bus.  Or it could display an error dialog.

(The worst thing to do is run but lose user configuration and data
from multiple processes doing uncoordinated writes to the filesystem,
which apparently some people think is acceptable, but I certainly
don't)

Realistically, DBus changing the "session" bus to user-machine doesn't
really make things significantly harder here for the environment.  I
think it's entirely acceptable for us to push this problem to
environment maintainers; they're the ones who should be in the
business of caring (or not caring) about multiple logins.

The main question I have is about *apps* - what's the story we tell
app authors who care about multiple logins?  Don't claim a bus name
then?  Seems reasonable to me.