04:17 <orowith2os[m]> hmm. Looking at the wl_output docs, it seems to imply that the done event gets sent after every event, or should it be sent when the global object there is all done being set up?

04:20 <orowith2os[m]> I see implementations sending the done event at the end, so I'm assuming it should be sent after the global is done being set up.

04:20 <orowith2os[m]> and then, I should send new events for any changes, also followed by done events for them?

04:22 <danieldg> yes

04:22 <orowith2os[m]> okay, I guess that's something that could maybe be done a bit better with in the protocol, then

04:24 <orowith2os[m]> the done event says it right, but the other events make it seem as if the done event is sent more than it really is

08:47 <wlb> wayland-protocols Merge request !389 opened by probonopd (probonopd) Draft: ext-toplevel-placement.v1: new protocol https://gitlab.freedesktop.org/wayland/wayland-protocols/-/merge_requests/389

13:45 <orowith2os[m]> riteo: how was your wayland proxy stuff coming along?

14:23 <emersion> zzag, pq, swick[m]: it seems like there's a contradiction in wp_image_description_info_v1.target_primaries

14:23 <emersion> it reads:

14:23 <emersion> > If target color volume is equal to the primary color volume, then this event is not sent.

14:23 <emersion> but the interface description says it's always sent

14:24 <emersion> which should it be?

14:43 <linkmauve> pq, I once wrote https://hg.linkmauve.fr/tablet-emu/ to emulate a tablet in order to test zwp_tablet_manager_v1, it can probably be simply changed to expose a touchscreen instead of a tablet, the evdev data is quite similar in both cases.

14:52 <linkmauve> Note that it works better if you have two outputs, put it on one, and configure the compositor to associate the tablet with the other one.

15:19 <swick[m]> emersion: would be best to track that as an issue

15:20 <swick[m]> there isn't an immediate right or wrong answer to it

15:33 <emersion> ah, i assumed there would be an immediate "we should delete this paragraph, we forgot about it when doing https://gitlab.freedesktop.org/swick/wayland-protocols/-/merge_requests/105#note_2749233 "

18:49 <riteo> Hi orowith2os[m], it's going along well I think. Still polishing stuff. Most importantly I improved global handling so that I could hook up a custom control interface for the first client, but I still gotta figure out the API as it has to play nice with Godot's own process embedding API

18:50 <riteo> It's still super-duper WIP but if you really need to see the prototype I can make a public branch in my repo ig.

18:51 <riteo> the code is now way more pretty than before

18:56 <orowith2os[m]> riteo: would definitely like to take a peek

19:08 <riteo> aight

23:13 <orowith2os[m]> whot: it wouldn't be an API break for me to go through libinput and change some pointers to `const`, right?

23:13 <orowith2os[m]> mainly thinking of libinput_config_accel_set_points

23:14 <orowith2os[m]> My mutter patches are casting it from a const pointer to raw pointer, and I'd rather not, so if libinput could do better there, it would be nice.