09:26 <riteo> hiiiii! It's me, that godot person! This time I'm having some issues with making wayland+x11 builds work on either platform without the "opposite" platform's libraries

09:27 <riteo> is this possible? I found a dload wrapper but while it works with X11 stuff it doesn't work well with wayland

09:28 <riteo> can I bundle the wayland client library into the client?

09:29 <vyivel> i'd use something like build-time flags for this tbh

09:30 <riteo> me too, but the other godot people want to make a single binary that can work with or without either library available

09:30 <riteo> they're already currently dloading dbus and fontconfig

09:30 <vyivel> i think the same should work for libwayland

09:31 <riteo> at least that wrapper generator I foudn has issues with parsing its types

09:31 <riteo> although I could make my own wrapper, couldn't I?

09:31 <riteo> wayland-client is pretty small IIRC

09:34 <vyivel> iirc firefox bundles libwayland-client stubs, loads libwayland on run-time and replaces(?) the stubs with actual functions if available

09:34 <vyivel> might be terribly misremembering that though

09:35 <riteo> is there a reason on why I can't bundle the whole wayland-client library?

09:35 <riteo> like, is it too heavy, platform dependant, or just a bad idea in general?

09:36 <vyivel> i think you can?

09:36 <vyivel> it's fairly small

09:36 <riteo> tbh I'd be tempted to just bundle it

09:36 <riteo> godot bundles way bigger libraries and Xlib gets wrapped nicely by that utility

09:36 <vyivel> you'd have to deal with updates on your own, but that's about it i presume

09:37 <riteo> mh, I'll consider this

09:37 <riteo> otherwise I think I'll either fix the wrapper thing or just make my own library wrapper, this thing's small after all

09:47 <jadahl> riteo: you can check how sdl does it (it does dlopen() and dlsym())

10:05 <riteo> sorry, I timed out

10:05 <riteo> jadahl: can I?

10:05 <riteo> wouldn't I have licensing problems?

10:20 <riteo> oh I think I fixed the wrapper generator. I think that'll do

10:20 <riteo> thanks for the help y'all!

12:04 <mntirc> soreau: apparently there is a patch now https://github.com/libffi/libffi/pull/745/commits/80d984377d8d5fc3b86786b8a99fdd8d2ee4ec4c

12:17 <JEEB> hi, I see there is KMS/DRM HDR related stuff in libweston, and I'd like to pick up someone's brains on this. if I look at the CTA specifications, I see that there's probably two things at play: a) set the colorimetry in the Aux Video Information InfoFrame to RGB, BT.2020 b) set the Transfer Function in the Dynamic Range and Mastering InfoFrame to PQ

12:17 <JEEB> as an example of BT.2020+PQ

12:18 <JEEB> yet looking at the great support matrix for DRM/KMS properties, the support matrix for various graphics drivers is kind of weird. https://drmdb.emersion.fr/properties/3233857728/Colorspace VS https://drmdb.emersion.fr/properties/3233857728/HDR_OUTPUT_METADATA

12:18 <JEEB> or do drivers just go lalala and configure the colorimetry flag "behind our backs" when the HDR output metadata is set?

12:19 <JEEB> to some value which is not really defined

14:37 <soreau> mntirc: ah, good

20:05 <zubzub> I'm trying to pinpoint some XWayland error on NVIDIA driver and tracked it to a call to gbm_bo_create_with_modifiers(..) inside xwayland-glamor-gbm.c (L295) which returns NULL but I have no idea *why* it returns NULL. format is set to 'GBM_FORMAT_XRGB8888' and modifiers set to: 72057594037927935 (whatever that is) :-/

20:05 <zubzub> cc ofourdan ^

20:21 <dottedmag> 72057594037927935 = (u64)-1

20:21 <dottedmag> doesn't sound sane

20:21 <JEEB> yea

20:22 <zubzub> ok so my compositor is probably advertising an illegal modifier

20:35 <zubzub> 72057594037927935 == DRM_FORMAT_RESERVED

20:48 <dottedmag> Oops, I didn't count the number of "ff"s :)

20:58 <zubzub> hmm my other modifiers all seem fine (valid? nvidia modifiers): 0x3000000004FE010, 0x3000000004FE011, 0x3000000004FE012 etc.

22:00 <zubzub> ooh cool I get the same XWayland error on Weston 9 as well

22:01 <zubzub> Guess I'm probably not doing anything wrong after all 🥲