Consolatis

Just a theory, not verified at all: This sounds like firefox is submitting an empty buffer (e.g. allocated but not actually written to). I am not sure who is responsible...

Does it happen as well when starting labwc with software rendering, e.g. `WLR_RENDERER=pixman labwc` ?

Just saw this, its possibly the same issue: https://github.com/swaywm/sway/pull/6844#issuecomment-1509344115 So starting labwc like `WLR_SCENE_DISABLE_VISIBILITY=1 labwc` might help.

With the new hint, it seems my earlier theory is wrong. What seems to actually be the case is that there is a bug in firefox which, during startup, claims...

> Move `handle_menu_keys()` to `menu/input.c` Do we really want to do that? It uses various `XKB_KEY_` defines and basically just maps keys to actions. I think `src/keyboard.c` is a good...

Marked the pipe menu entry as done. Is there anything you want to see in 0.7.2? Otherwise I'd suggest that we move this to the 0.7.3 milestone.

We don't currently support mouse emulation for touchscreen devices. If the application supports the wayland touch protocol it should work for them though, e.g. gestures and friends.

I assume you have a single display? What version of labwc are you using, somewhat recent master branch or one of the releases?

Yes, the fullscreen state is indeed reset when turning the monitor off / disabling it. This was a deliberate choice because usually the windows will be transferred to a still...

So something like this (not even compile tested) ```c /* In output.c */ size_t output_get_active_count(struct server *server) { size_t count = 0; struct output *output; wl_list_for_each(output, &server->outputs, link) { if...