| 5 Jan 2025 |
 @hexa:lossy.network | init at 1.4.1 | 03:42:22 |
| @hexa:lossy.network | common clause was put in place in version 0.10.0 | 03:43:26 |
| @hexa:lossy.network | https://github.com/NixOS/nixpkgs/pull/371053 | 03:48:33 |
|  Xiyan (Sesquipedalio) joined the room. | 07:16:37 |
 ibizaman | Does everyone need to build triton-llvm locally these days or do I have some issue in my config? The approximative commit that still allows me to pull from the cache is https://github.com/NixOS/nixpkgs/commit/4a35f2b7eb8a8ad8ca54fd1358d8805b620af9b2, then 1 day afterwards until today any commit I tried requires me to build it.
The dependency of why I even need triton-llvm in the first place is (from nom):
┃ │ │ ┌─ ⏵ triton-llvm-19.1.0-rc1 (buildPhase) ⏱ 1m35s
┃ │ │ ┌─ ⏸ python3.12-triton-3.1.0
┃ │ │ ┌─ ⏸ python3.12-torch-2.5.1
┃ │ │ ┌─ ⏸ python3.12-safetensors-0.4.5
┃ │ │ ┌─ ⏸ python3.12-transformers-4.47.1
┃ │ │ ┌─ ⏸ python3.12-ctranslate2-4.5.0
┃ │ │ ┌─ ⏸ python3.12-faster-whisper-1.1.0
┃ │ │ ┌─ ⏸ wyoming-faster-whisper-2.2.0
┃ │ ├─ ⏸ unit-wyoming-faster-whisper-tiny-fr.service
┃ ├─ ⏸ system-units
┃ ┌─ ⏸ etc
┃ ⏸ nixos-system-baryum-25.05pre-git
I'm happy to bisect the commits more if needed.
| 10:23:53 |
| @hexa:lossy.network | There was a broken test | 10:57:46 |
| ibizaman | This, right? https://github.com/NixOS/nixpkgs/commit/b4240209aae34579d2e8e3de3203285d0c005f22
I tried that earlier but it still required me to build triton-llvm. But now it works!
That makes me think, is there a way to know beforehand what is in the NixOS cache? | 11:26:55 |
| ibizaman | * This is the fix, right? https://github.com/NixOS/nixpkgs/commit/b4240209aae34579d2e8e3de3203285d0c005f22
I tried that earlier but it still required me to build triton-llvm. But now it works!
That makes me think, is there a way to know beforehand what is in the NixOS cache? | 11:29:38 |
| ibizaman | * This is the fix, right? https://github.com/NixOS/nixpkgs/commit/b4240209aae34579d2e8e3de3203285d0c005f22
I tried that earlier but it still required me to build triton-llvm. But now it works!
That makes me think, is there a way to know beforehand what is in the NixOS cache? Like at least something saying this package in this nixpkgs commit is available in the cache? | 11:32:34 |
 CRTified | In reply to @hexa:lossy.network
and we misrepresent it as agpl3.0-only since basically forever Ouch | 11:39:57 |
| CRTified | In reply to @hexa:lossy.network
CRTified: urgh, tandoor-recipes uses the common clause, which makes it non-free software Probably because they also run tandoor as a paid saas. But the README.md section about licensing confuses me | 11:47:20 |
| CRTified | It seems to mix common cause and agpl?! | 11:47:36 |
| @hexa:lossy.network | it doe | 11:49:06 |
| @hexa:lossy.network | * it does | 11:49:07 |
 K900 | python3.13-adb-shell tests are failing for me on master :( | 14:15:01 |
| K900 | (aarch64) | 14:15:03 |
| @hexa:lossy.network | I know | 14:19:37 |
| @hexa:lossy.network | sorry, a few things are broken on 3.13 | 14:19:47 |
| @hexa:lossy.network | async generator issue | 14:19:52 |
| @hexa:lossy.network | started here https://github.com/JeffLIrion/adb_shell/pull/238 | 14:20:24 |
| @hexa:lossy.network | if you want to look into that, note that this piece of shit silences warnings | 14:30:06 |
| @hexa:lossy.network | https://github.com/JeffLIrion/adb_shell/blob/master/tests/async_wrapper.py#L10-L17 | 14:30:06 |
| K900 | In reply to @hexa:lossy.network
sorry, a few things are broken on 3.13 I'll look into it maybe when I'm back from the dentist | 14:31:21 |
| @hexa:lossy.network | most likely we can just yeet the tests on 3.13. | 14:42:47 |
 Tako | In reply to@hexa:lossy.network
home-assistant 2025.1.0 merged just noticed 24.11 is still on 2024.11.1, will it get bumped to 2025.1.0 once it stabilizes? | 15:39:29 |
| K900 | In reply to @hexa:lossy.network
most likely we can just yeet the tests on 3.13. On my way home now | 16:00:00 |
| K900 | Ready to yeet | 16:00:07 |
| @hexa:lossy.network | never | 16:22:46 |
| @hexa:lossy.network | * we never backport home-assistant versions | 16:22:56 |
| @hexa:lossy.network | too disruptive for the whole python package set | 16:23:14 |
