layer_shell: destroy layer_surface on assigned output destruction

[llyyr]

According to the spec, the closed event should be sent when the surface is no longer shown, because the output may have been destroyed or the user may have asked for it to be removed. In such cases, the clients should destroy the resource.

This fixes mako not being able to show notifications if the assigned output was destroyed while a notificataion was still visible

[llyyr]

I can't really understand what the previous code was trying to do here, so I don't know if this is completely incorrect but this behavior matches what every other compositor does with on the output disable event. Even sway before the scene port, so probably just a tab complete fail that went unnoticed?

[emersion]

It seems like wlr_layer_surface_v1_destroy() was called on output destroy before 188811f.

I think we're missing a layer shell surface destroy handler? Here the struct sway_layer_surface seems to be leaked when wlr_layer_surface_v1_destroy() is called? I'm actually surprised we don't hit asserts due to dangling listeners.

(There are other cases where a layer surface might get destroyed, e.g. when a client disconnects.)

[llyyr]

I think we're missing a layer shell surface destroy handler?

Seems like this is handled by handle_node_destroy, including clearing out dangling listeners. I modified my patch destroy the layer surface in handle_node_destroy instead if we don't have any outputs

[emersion]

I think this still leaves the struct sway_layer_surface behind when a client destroys the layer shell surface.

[llyyr]

I think this still leaves the struct sway_layer_surface behind when a client destroys the layer shell surface.

handle_node_destroy is notified when a client destroys the layer shell surface, and we free struct sway_layer_surface there. It's a bit confusing that we're using wlr_scene_node's destroy event like this, but I think it makes sense

[emersion]

handle_node_destroy is notified when a client destroys the layer shell surface

What is the mechanism destroying the node when the layer shell surface is destroyed?

[llyyr]

It's handled here https://gitlab.freedesktop.org/wlroots/wlroots/-/blob/master/types/scene/layer_shell_v1.c#L15-20

[emersion]

Oh right, my bad, I forgot that we were using the scene-graph layer shell helper now…

I think I preferred the previous version of this patch.

[llyyr]

Changed it back

[emersion]

Thank you, and sorry for the noise!