Handling of layout/input methods in Unity: what should we do for the LTS?

[Sebastien Bacher]

Le 13/11/2013 17:49, Loïc Minier a écrit :
> How will the Mir/Wayland approach look like?  Could we leverage it for
> Unity 7?
I expect the approach in "the new world" to be similar to the recent 
work. It should be easier under Mir/Wayland since we are not going to 
have to fight with X grabs and xkb limitations there (e.g Mir should be 
watching for those events and sending the signal to do the action 
associated to the keybinding).

That's a topic that should be discussed with the Mir/Unity8 though ... 
do you know if that's on the roadmap for this cycle (handling of 
external keyboards and layouts/input methods/keybindings on the desktop)?

> unless it's
> a lot of work to rollback, in comparison to moving to the new world.
Well, we are almost done with the work we have been doing. The issue is 
that we are hitting the problems I listed, especially the ones with apps 
like libreoffice that require to fix the applications. I expect 
applications are not going to be an issue in touch because there is less 
"legacy" and the ones newly written or using the standard toolkits 
shouldn't have those issues.

Rollback is not that much work in any case.
>  From a risk point of view, the second option seems like the safest one,
> but from an efforts point of view this depends on how much work it is to
> produce an Unity 8 image with a new keyboard layouts / shortcuts
> solution + finding a solution for Ubuntu Desktop.
The solutions are going to be different in both setups (since the input 
stack used is not the same), it also depends on the input method 
framework we are going to use (there is still the ongoing ibus or fcitx 
discussion, with a session at vUDS next week)

Cheers,
Sebastien Bacher