http://bugs.winehq.org/show_bug.cgi?id=22881
--- Comment #1 from Jerome Leclanche adys.wh@gmail.com 2010-07-30 22:30:43 ---
From #winehackers today
[04:13:44] <madewokherd> it's highly unlikely that a window manager that goes to the trouble of leaving windows undecorated when they request it won't also move the windows when asked [04:13:53] <madewokherd> that would produce a very bad experience [04:14:21] <madewokherd> oh, unless it just doesn't support _NET_WM_MOVERESIZE [04:14:46] <madewokherd> however [04:14:51] <Adys> Would explain why metacity complains [04:15:15] <madewokherd> oh [04:15:18] <madewokherd> we check for it and do nothing [04:15:36] <madewokherd> as opposed to checking for it and falling back to another implementation ... [04:20:32] <madewokherd> we apparently had a more user32-oriented solution [04:20:37] <madewokherd> back in 2001 [04:22:27] <madewokherd> it probably didn't work very well [04:22:38] <madewokherd> but this sort of thing should still be there, for the unmanaged case at least [04:23:11] <madewokherd> if (!data->whole_window || !data->managed || !data->mapped) return -1; [04:23:16] <madewokherd> and now it makes sense [04:23:38] <madewokherd> "return -1" means "use a default implementation" oslt [04:23:41] <madewokherd> not "do nothing" [04:24:14] <madewokherd> Adys: so yeah, that should really work, and more investigation is needed to find out why it doesn't in openbox