07:18 <fleischie> scorpion2185[m]: https://wiki.archlinux.org/title/NVIDIA#DRM_kernel_mode_setting it says to "add the `nvidia_drm.modeset=1` kernel parameter" which depends a bit on your boot manager.

13:24 <novakane> hmm #771 is weird, I tried and indeed river doesn't send focused_view event anymore when changing Firefox tab

13:24 <novakane> but that doesn't really makes sense that the update to zig 0.10 broke that

13:28 <ifreund> novakane: did you try before and after that commit to make sure it was that one that caused the regression?

13:29 <ifreund> I was meaning to but haven't yet. Definitely weird

13:29 <novakane> yeah I tried with both

13:32 <ifreund> very strange, hope that's not a miscompilation somewhere

13:33 <ifreund> My next step would be adding alot of print statements along the code path that's supposed to trigger that

13:33 <ifreund> I'm currently still looking into that Xwayland PR I've been wanting to land for a while

13:34 <novakane> yeah I'll try to find more info, I quickly wrote something to get Wayland debug message sent by river status protocol

13:35 <novakane> well without using Wayland debug when launching river I mean, that's too noisy

13:43 <ifreund> yeah for sure

14:17 <novakane> so clearly SeatStatus.sendFocusedView() is no more called when switching tab

14:17 <novakane> now I need to find where that stop working exactly

14:56 <novakane> it seems to never enter this loop for some reasons https://github.com/riverwm/river/blob/master/river/View.zig#L563-L565

14:57 <leon-p> ifreund: btw, what do you think of additionally dumping warn and err messages to syslog? makes it possible to get at least some info on what went wrong when river weirds out and you are not logging it actively

14:57 <leon-p> I used that recently to find out that river was OOMing on me, for example

15:00 <ifreund> leon-p: I'd be happy for river to log warn and higher to syslog as well by default. Let's also add a -no-syslog option though

15:01 <leon-p> sure!

15:02 <ifreund> also, river was OOMing on you? did you disable overcommit or something?

15:06 <novakane> hmm that the if condition just above that is never true instead it seems

15:07 <novakane> Why dos this stop to be true after the zig update though, I have no idea

15:07 <novakane> s/dos/did

15:20 <ifreund> novakane: it's likely a compiler bug, I'd try restructuring the code to put as much as possible on separate lines and see if that changes anything

15:21 <ifreund> if it does, submit an upstream zig bug report

15:21 <novakane> ifreund: it is `seat_node.data.focused.view == &self` that is false now

15:22 <ifreund> novakane: oh I see the problem

15:22 <tleydxdy[m]> hmm have anyone set up zig 0.10 on arch?

15:22 <ifreund> I bet this fixes it https://0x0.st/o7VM.diff

15:22 <tleydxdy[m]> I'm hitting lots of compile errors

15:22 <ifreund> if you confirm please do open a PR

15:23 <ifreund> tleydxdy[m]: I'd recommend just using a statically linked zig tarball from https://ziglang.org/download

15:23 <ifreund> at least until someone gets around to packaging it.

15:23 <tleydxdy[m]> it is in staging apparently

15:23 <tleydxdy[m]> I got the pkgbuild from their repo

15:24 <tleydxdy[m]> but src/zig_clang.cpp all the asserts fail in here

15:24 <tleydxdy[m]> I need to use llvm 15 correct?

15:24 <ifreund> novakane: sorry, should be https://0x0.st/o7Vu.diff

15:24 <ifreund> tleydxdy[m]: yeah, llvm 15 is required to build zig 0.10.0

15:26 <tleydxdy[m]> aha, my clang is still on 14

15:27 <novakane> ifreund: yeah I already fixed the diff, no worries, and it seems to fix it \o/

15:27 <ifreund> sweet!

15:28 <ifreund> not a miscompilation then, a latent bug in river

15:38 <novakane> luckily you found that fix quickly because I would probably have looked at that &self all day :P

15:39 <novakane> I don't have waybar though so if someone want to test it

17:37 <scorpion2185[m]> novakane test what?

17:39 <scorpion2185[m]> "river/window": {... (full message at <https://libera.ems.host/_matrix/media/v3/download/libera.chat/eebf29dbb77c7956414e77d8a5c3b385e7ba0a97>)

17:41 <leon-p> ifreund: it was OOMing because my workstation shares memory between CPU and GPU. So when I was doing something graphically intensive (some python image processing library), something went wrong

20:06 <fleischie> /quit

20:06 <fleischie> Whoops. :O

22:09 <talismanick> I haven't used Zig before, but I have used C and C++. Not at all familiar with the details of wlroots or Wayland itself. If I wanted to help along the tablet-v2 bindings (so I could use my pen), where should I begin?