[08:42] <anpok> alan_g: tested your branch.. now looking at other window type placements..
[08:42] <anpok> +it works..
[08:43] <anpok> for a tooltip .. the aux_rect() window specification parameter should be filled, shoudlnt it?
[08:43] <anpok> I think that case is still wrong in the gdk backend.
[08:46] <alan_g> @"for a tooltip" the aux_rect() should be used for placement
[08:46] <alan_g> DId I answer your "needs info"?
[09:02] <anpok> alan_g: I think what do not understand here is why window_info needs to be updated.. Is the WindowInfo object destroyed when the Window is hidden?
[09:03] <anpok> Or is WindowInfo kept alive but not updated for some reason when the parent Window is moved around?
[09:04] <alan_g> anpok: no... the restore_rect is only updated on state transitions.
[09:05] <anpok> ah!
[09:05] <alan_g> I think it will be clearer when I update the follow-up branch to fix other issues I found as a result.
[09:05] <anpok> Now.. ok
[09:06] <anpok> I thought you would be moving the location of WindowInfo to match the one of Window.. But instead you are preparing WindowInfo for the 'restore' step
[09:06] <alan_g> Yep
[09:29] <sil2100> alf: ping
[09:31] <duflu> sil2100: Paternity leave :)
[09:32] <sil2100> duflu: argh, damn, since he's the only admin of the new repowerd team ;/
[09:32] <sil2100> duflu: thanks for the info
[09:32] <sil2100> duflu: do you know till when?
[09:32] <duflu> Of course. That's how these things always go
[09:32] <duflu> No I don't know
[09:36] <alan_g> anpok: while you still remember gtk-mir... When running (e.g.) bin/miral-kiosk --startup gnome-terminal the terminal doesn't paint the full (fullscreen) surface. Is the fix trivial?
[10:01] <anpok> hm miral-kiosk does not start up for me
[10:02] <alan_g> Odd. Get an error?
[10:02] <anpok> ah ok .. the startup or startup-apps does not work
[10:02] <anpok> http://pastebin.ubuntu.com/19259936/
[10:04] <anpok> now it fails with: Error creating terminal: No screen 0 on display "Mir"
[10:05] <alan_g> anpok: the pastebin looks like trying to start mesa-kms without root
[10:05] <alan_g> Which is the same code that all the other Mir servers run. Kiosk isn't any different there
[10:06] <anpok> so I did not manage to get gnome-terminal to start
[10:07] <alan_g> anpok: gnome-terminal doesn't work as root (without fiddling). try running on X11
[10:08] <alan_g> Or... hack scripts/testrun to run kiosk
[10:13] <anpok> oh i get a bunch of errors with radeonsi now..
[10:14] <anpok> so couldnt see it in action yet..
[10:14] <alan_g> nm was just a passing thought
[10:16] <alan_g> anpok: another thought: did you fix this in your gtk-mir travels? lp:1445543
[11:23] <bregma> sil2100, alf is gone at least two weeks, supposed to be back 2016.08.02
[11:24] <bregma> sil2100, I'm sure canonical IS can probably do something if there's an emergency....
[12:00] <anpok> alan_g: something worth trying
[12:20] <anpok> alan_g: still exists.. the input event reaches gdk.. from then on it is a bit unclear.
[12:57] <alan_g> anpok: thanks for checking
[12:59] <anpok> hmm or actually I did not check whether the event reaches the right surface..
[12:59] <anpok> Now looking at a ci issue with the nested input device state event tests I added recently..
[13:02] <anpok> alan_g: the way that window looks it seems to not receive keyboard focus.. maybe it got mapped to a window type not allowed to receive user input?
[13:03] <anpok> alan_g: do we have something in scene report showing that?
[13:03] <alan_g> anpok: I suspect not (yet)