> As for Gerry's suggestion of removing this code completely and seeing if
> anything breaks yet, I believe this code is still used. "role" is only
> apparently used today if we are the OSK role, but that's in Mir code. So
> clearly Mir is reserving the right to care about the role we set.
>
> As for the session property, that does seem to percolate down through to
> the Android layer and affect the oom scores and similar issues.
So I dug into this, and you're right, it's used still. Not for long however as QtComp makes this unnecessary.
> As for Gerry's suggestion of removing this code completely and seeing if
> anything breaks yet, I believe this code is still used. "role" is only
> apparently used today if we are the OSK role, but that's in Mir code. So
> clearly Mir is reserving the right to care about the role we set.
>
> As for the session property, that does seem to percolate down through to
> the Android layer and affect the oom scores and similar issues.
So I dug into this, and you're right, it's used still. Not for long however as QtComp makes this unnecessary.