| 29 Jan 2023 |
 Arson Copperflame | That looks promising! Do you know if that'll work with Intel iGPUs as well? | 07:55:59 |
 K900 | I couldn't get it to work on Intel because the driver requires some weird sonames we don't have in nixpkgs | 07:56:26 |
| K900 | I'm not sure why but they just dlopen a bunch of stuff | 07:56:49 |
| K900 | And it seems like all the vendors do that | 07:57:06 |
| K900 | But at least for AMD it's just openssl | 07:57:30 |
| K900 | Which we have | 07:57:32 |
| K900 | I haven't really looked into Nvidia because my only Nvidia system is an Optimus setup, and it tries to enumerate the Intel chip first | 07:58:36 |
| K900 | And that | 07:58:45 |
| K900 | * And that barfs | 07:58:47 |
| K900 | Uhh | 15:36:38 |
| K900 | ~
❯ cat /run/tmpfiles.d/x11.conf
# Note: This file is generated by WSL to prevent systemd-tmpfiles from removing /tmp/.X11-unix during boot.
| 15:36:40 |
| K900 | wot | 15:36:41 |
| K900 | I guess they don't link it with the systemd rule anymore | 15:38:59 |
| 30 Jan 2023 |
| K900 | https://github.com/nix-community/NixOS-WSL/commit/bbfe6ad9b877935ee74c98030193e557bcfed354 | 05:58:22 |
| K900 | Ugh | 05:58:23 |
| K900 | Had to push directly to master | 05:58:27 |
| K900 | And another followup fix: https://github.com/nix-community/NixOS-WSL/pull/208 | 05:59:14 |
| Arson Copperflame | In reply to @k900:0upti.me
https://github.com/nix-community/NixOS-WSL/commit/bbfe6ad9b877935ee74c98030193e557bcfed354 Both the default config as well as my config change that, therefore I forgot to update it here. So that one's my bad | 09:04:07 |
| K900 | At least mine didn't so I spotted it immediately | 09:07:34 |
 Sandro 🐧 | In reply to @k900:0upti.me
sent an image. I could test it on wsl but I don't have that much time tbh | 10:48:56 |
| K900 | Maybe wait for the staging merge then | 10:49:24 |
| K900 | That should have dzn by default | 10:50:09 |
| K900 | You'll still need to figure out the right LD_PRELOAD incantation though | 10:50:21 |
| Sandro 🐧 | K900: maybe something little to make a little bit nicer
▶ RUST_BACKTRACE=full /bin/init
thread 'main' panicked at 'Failed to set up logger...: Io(Os { code: 13, kind: PermissionDenied, message: "Permission denied" })', src/main.rs:126:21
stack backtrace:
0: rust_begin_unwind
1: core::panicking::panic_fmt
2: core::result::unwrap_failed
3: nixos_wsl_native_systemd_shim::main
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.
| 13:53:25 |
| K900 | Oh wow how did you even do that | 13:53:43 |
| K900 | Or do you just mean running the shim with RUST_BACKTRACE=full in general? | 13:54:37 |
| K900 | That's not going to be very useful, we don't ever panic | 13:54:56 |
| K900 | All the errors bubble up to main and THEN panic | 13:55:13 |
| K900 | So the backtrace will look exactly like this every time | 13:55:23 |
| K900 | But the error object itself will preserve context | 13:55:35 |
