08:41 <pq> vanfanel...

08:42 <pq> I have no idea how that part of the code actually works.

15:55 <Vanfanel> Sorry to insist, but... On Weston, in desktop-shell/shell.c, what would be the correct way to simulate a click on a window when it goes fullscreen? Is set_fullscreen() a good place to do so?

15:57 <pq> I have no idea how that part of the code actually works.

15:57 <jadahl> why do you want to simulate a click?

15:58 <Vanfanel> pq: thanks, I will keep trying :)

15:58 <pq> because weston requires a click to grant pointer confinement, but a fullscreen surface becoming active should not need a click

15:58 <Vanfanel> jadahl: what pq said ^

16:00 <pq> Vanfanel, I think there should be two triggers: becoming fullscreen while active, and becoming active while fullscreen.

16:00 <pq> either one should be enough to grant confinement

16:00 <pq> also maximized, I suppose

16:01 <pq> also, it's not simulating a click really, it's just rigging the state tracking so that the input machinery grants confinement when asked

16:02 <pq> the obstacle here is that the input machinery has no concept of "fullscreen" or even top-level

16:03 <pq> hence the alternative route of pretending the window was clicked

16:06 <Vanfanel> pq: so I should be changing maybe_enable_pointer_constraint() as a definitive solution too? Since that's where pointer confinement is evaluated... Maybe use a flag from somewhere... but from where? Isn't set_fullscreen() where the window is set fullscreen?

16:09 <Vanfanel> pq: there's also the problem that I don't know how to retrieve the view from maybe_enable_pointer_constraint(): if you look at it, it iterates over all views of a surface and choses the one that was clicked.. but no one was clicked, so I should add a "is_fullscreen" flag from somewhere to the view that was just made fullscreen. The problem then is: WHERE is the right place to mark a view that is being

16:09 <Vanfanel> made fullscreen? Is set_fullscreen() the place to do so?