01:19 hays: wow. I think for mali, I need to be using the -gbm version of the driver and not the wayland-gbm version
01:20 hays: very counterintuitive but the diff in symbol tables here https://h0c.us/v/o6TzQw/r I think these functions are almost all provided by libwayland-egl, libwayland-client, or libwayland-server.
01:23 hays: and that also makes sense. why else 'clean up' the interface where wayland-egl.so is now provided by wayland if you don't get to use a generic binary that doesn't have knowledge of wayland
10:24 Vanfanel: jadahl: Ok, understood. Is there a way to know if a view is fullscreen? I could then modify maybe_enable_pointer_constraint() to grant pointer confinement to surfaces with a fullscreen view!
10:25 Vanfanel: jadahl: If there's no fullscreen property on weston_view (It doesn't seem to have one), then, where are the weston_views made fullscreen, please¿
10:27 daniels: hays: it looks like your symbol table dumps include all the symbols that are used, not just the ones that are defined. you definitely need the wayland-gbm build.
14:06 hays: daniels: sorry, i don't follow. the above was a diff
14:37 hays: you are right--its even more striking if I use nm -gD --defined-only
14:37 hays: https://h0c.us/v/VSL5gA
14:58 hays: fwiw i was able to link wayland/wlroots/sway/cairo with the -gbm version... still need to test it though
15:33 hays: to me this makes sense logically--but definitely need to confirm, and interested to hear about reasons why it wont or can't work
16:02 hays: wl_egl_window_retain and wl_egl_window_release are missing---but these don't seem to be actually used or defined publicly
16:09 jadahl: Vanfanel: "fullsrceen" is a window management level property of weston_desktop_surface. for a view that *may* mean it simply covers an output, but e.g. a logically fullscreen surface might e.g. animate towards being effectively fullscreen, and during said animation, the view isn't technically fullscreen
18:48 hays: daniels: this is how im thinking about it-- not sure I am right but wanted to communicate the idea -- https://file.st5ve.com/image.png
20:14 Vanfanel: jadahl: So, in your view, what should I look at in order to grant pointer confinement to a surface? I mean: what property of what view or what property of the surface itself can tell me "yes, this is NOW a fullscreen surface: it can have pointer confinement".