On 23.10.2014 5:12, Wang, Quanxian wrote:
[Wang, Quanxian] No one against the matured multi-user mechanism. We just separate multiple displays from multiple users. Your design is to bind them together.
No, the default configuration binds seat to a session which is in turn bound to a user. Seat can consist of one or more displays.
Or you can have multiple logical seats on one physical seat like I have now in my own test configuration.
There is no hard binding between physical and logical seats. But systemd/logind considers binding between physical seat, session and user.
[Wang, Quanxian] Still, multiple users is not multiple seat/display. Multiple seat/display is not private thing. They could be used by any user account or apps or other resource. It should be out scope of multiple user design. You can use it, but should not be design under the theory of multiple user. Any user could select any display or any number of displays. Any apps could select, any other resource could select that if they want. Of course, suppose permission is allowed.
Multihead is already supported by the system by default. Multi-seat is not, that's why we've been modifying wayland/weston to support multi-seat.
It is privacy thing when you have have multiple simultaneous users. You need to isolate the users and their data. Including things like contacts synced from a phone.
App started by user A and it's data must stay within user A and App started by user B and it's data must stay within user B. Even if they would attempt to access each others data, input methods or display, it must not succeed.
Please draw a diagram how seats, sessions and users (UIDs) are linked in your mind. (seat is collection of displays, input methods, audio devices and other hardware, like USB connectors)
We follow the same seat/session/user model as systemd/logind, udev, weston, libinput, etc. use. Otherwise you'd need to modify all those...
_______________________________________________ IVI mailing list [email protected] https://lists.tizen.org/listinfo/ivi
