[systemd-devel] Someone working on gui for multiseat?

[Stef Bon]

2013/1/25 Olav Vitters <olav at vitters.nl>

> > So perhaps it would make sense to look at implementing the GUI as a
> > component that will eventually be part of the GNOME systemsettings? i.e.
> > make a bold statement about where it should live. If it proves to be
> > useful and popular, then some KDE/Qt centric folks may port it to their
> > system settings GUI too.
>
> For gnome-control-center (3.7.x name being 'Settings' :P), best to check
> beforehand where it would fit and how it should interact and so on.
> There is an Wacom tablet thing in there. Ideally it should only be shown
> when needed (e.g. when it finds hardware), etc. Only something in
> Settings is not always enough, lately goal is providing the best
> experience, not just making something work. E.g. if it can be automatic,
> do it automatically. If it really needs something manual, maybe ask when
> there the need to ask arises, etc.
>
> For above and more, best to check beforehand. Assume that'll result in
> some API that multiple DE can use.
>
>
Thanks a lot for you remarks. It's indeed my intention to get some API
first.

First I want to say that the current settings for the display(s) is very
good. I'm working with two monitors I want to act as one big display, and
it's very easy to make this work.

analyse things first:

1. When dealing with a docking station (like those from  plugable) on a
laptop, the reason the user wants to use it, is to use a bigger screen and
a "real" keyboard and mouse. In this case you don't want a second session
(with greeter etc), but that the current session is at the same time
available on the docking station (and possible not available anymore on the
native screen).

This would be the default behaviour. The "other" behaviour (a second
session) is done when the user (administrator) has configured it.

Above choice is only done for the first docking station. When another
device is plugged in (so there will be more than one docking station
plugged in) then another session with login is always started.

This also counts for the (new?) possibility to connect such a docking
station to your (smart) phone and tablet. (Ubuntu has announced that it's
working on making it work on smart phones and tablets, I guess it's also
good keep this in mind for Fedora)

I think that it would be nice that you can set this behaviour somewhere
--before-- the device is plugged in.

2. When dealing with a desktop,  it depends on: is there already a "seat0":
eg  the "normal" seat using the standard monitor(s) and keyboard and
pointer device.

When it's present and working, then the default behaviour is: for every
docking station plugged in a new session is started.

When there is no seat0, then the first docking station becomes seat0,
simular to the laptop/smartphone case above.


So when looking at this:

a. for users in the admin group there should be somewhere in the settings
available the choice what to do when a docking station is plugged in:
"overtake seat0" or a new session. With defaults above described.

b. when the choice is "a new session"  a "set behaviour extra seats" dialog
appears in the settings only for user on the seat0 if he/she is part of
admin group

c. when the choice is "overtake seat0" the "set behaviour extra seats"
dialog will only appear after the first docking station is plugged in.

d. in the menu, somewhere at settings, an advanced dialog is to construct
"custom" seats, assigning hardware manually to construct a seat (when not
dealing with seat extenders/docking stations like those from plugable)


Does this look ok?

Stef Bon
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/systemd-devel/attachments/20130125/56cafdd3/attachment.html>