| 13 Nov 2024 |
 hexa (UTC+1) |
64m copying path '/nix/store/gmxz8f4b8pcqkwahz62jd3p37lj38qsb-python3.11-tensorflow-gpu-2.13.0' from 'https://nix-community.cachix.org'...
| 00:47:42 |
| hexa (UTC+1) | š¤ | 00:47:45 |
| hexa (UTC+1) | fetching via wget works with 30MB/s | 01:11:04 |
| hexa (UTC+1) | ok, broken nix-daemon š | 01:16:39 |
 GaƩtan Lepage | In reply to @hexa:lossy.network
next step is to test whether disabling the problematic test helps I tried adding pytest-timeout and --timeout=100 and the test suit seems to pass just fine. | 12:14:05 |
| GaƩtan Lepage | Without the need to disable any test. | 12:14:13 |
| GaƩtan Lepage | I thought that pytest-timeout would fail a test that is taking too long | 12:14:35 |
 SomeoneSerge (back on matrix) | In reply to @snektron:matrix.org
Is there a reason why https://github.com/NixOS/nixpkgs/pull/291471 is not merged? Hi, could you test https://github.com/NixOS/nixpkgs/pull/355470 | 13:52:32 |
| hexa (UTC+1) | In reply to @glepage:matrix.org
I thought that pytest-timeout would fail a test that is taking too long that is the idea | 17:08:18 |
| hexa (UTC+1) | but if no test gets stuck long enough, then it does nothing | 17:08:34 |
| 14 Nov 2024 |
 Snektron | In reply to @ss:someonex.net
Hi, could you test https://github.com/NixOS/nixpkgs/pull/355470 Sure, wow, thanks for working on that | 07:03:34 |
| Snektron | I'll test it when i get to a computer :) | 07:03:53 |
| Snektron | I need to build many things | 12:02:30 |
|  Alexandros Liarokapis joined the room. | 12:47:10 |
| Alexandros Liarokapis | hi all! is https://docs.nvidia.com/vpi/2.0/index.html packaged anywhere? | 12:47:51 |
| Alexandros Liarokapis | not sure if relevant to the team in which case sorry for the noise | 12:48:23 |
| SomeoneSerge (back on matrix) | In reply to @aliarokapis:matrix.org
hi all! is https://docs.nvidia.com/vpi/2.0/index.html packaged anywhere? Yes it's relevant, no it's not packaged afaik | 13:51:30 |
| SomeoneSerge (back on matrix) | In reply to @aliarokapis:matrix.org
hi all! is https://docs.nvidia.com/vpi/2.0/index.html packaged anywhere? * Hi! Yes it's relevant, no it's not packaged afaik | 13:51:37 |
| Alexandros Liarokapis | Thanks, I have no idea about its dependency graph, I will try to package it during the weekend | 14:15:47 |
| Alexandros Liarokapis | I suspect it's going to he painful | 14:16:14 |
| hexa (UTC+1) | does anyone have capacity to look into jax numpy 2 support? | 16:24:00 |
| hexa (UTC+1) | now that staging is post-branchoff we are going to make the switch for the package set, because it was demanded multiple times already | 16:24:34 |
| hexa (UTC+1) | I did post to the latest jax bump PR fwiw | 16:24:46 |
| GaƩtan Lepage | Unfortunately, I have neither the skills nor time to deal with that now... We are stuck at 0.4.28 because upstream has bazeled TM their build process and our derivation is completely broken.
Maybe Samuel Ainsworth or SomeoneSerge (utc+3) will know better. | 16:36:34 |
| GaƩtan Lepage | Sad situation... similar to the tensorflow one, although zeuner is working on this one. | 16:36:55 |
| SomeoneSerge (back on matrix) | i haven't even had time to review zeuner's tensorflow prs... | 16:45:41 |
| SomeoneSerge (back on matrix) | * i haven't even had time to review zteuner's tensorflow prs... | 16:45:52 |
| SomeoneSerge (back on matrix) | * i haven't even had time to review zeuner's tensorflow prs... | 16:46:03 |
 connor (burnt/out) (UTC-8) | In reply to @aliarokapis:matrix.org
hi all! is https://docs.nvidia.com/vpi/2.0/index.html packaged anywhere? as far as I can tell, it's only available via debian installers. I'd take a look at https://github.com/anduril/jetpack-nixos since they have some tooling set up to repackage stuff like that (they also already have VPI iirc, but maybe not a new version) | 17:40:58 |
| 15 Nov 2024 |
| Alexandros Liarokapis | I wish I could run nixos modules inside a light container or something on as non-nixos machine. Waiting for systemd-nspawn to make this easier. | 09:46:58 |
