| 25 Sep 2025 |
 connor (burnt/out) (UTC-8) | To clarify, are you building PyTorch and Triton with Nix and producing a container, or is this something else? Iād also like to see a reproducer of you have one handy | 15:07:20 |
| connor (burnt/out) (UTC-8) | here be the path to (more) madness | 15:08:12 |
 apyh | In reply to @connorbaker:matrix.org
To clarify, are you building PyTorch and Triton with Nix and producing a container, or is this something else? Iād also like to see a reproducer of you have one handy I'm building a container that includes PyTorch as a runtime dependency, and that PyTorch code calls torch.compile. Let me try to create a minimal repro :) | 15:30:19 |
| apyh | https://github.com/arilotter/nix-torch-compile
a minimal repro ! | 16:31:15 |
| apyh | this includes both things needed to run torch.compile inside a docker container | 16:31:30 |
 hexa (UTC+1) | flox stable nixpkgs is still stuck on aug 27th https://github.com/flox/nixpkgs | 16:50:13 |
| hexa (UTC+1) | not sure who can effectively poke them | 16:50:23 |
 Robbie Buxton | @apyh pr is here https://github.com/NixOS/nixpkgs/pull/446105 | 17:36:25 |
 SomeoneSerge (back on matrix) | Yes, I'm really embarrassed to say, but python3Packages.tensorflow has been defaulting to tensorlfow-bin for about a year now... and from what I've heard, even that is a PITA during upgrades | 23:30:50 |
| SomeoneSerge (back on matrix) | Soooo turns out flox-stable has nothing to do with nixos-stable | 23:31:25 |
| hexa (UTC+1) | so what do they build for nixos-stable then? | 23:32:29 |
| SomeoneSerge (back on matrix) | It's a tad confusing, but uhh they follow nixos-unstable, and when it passes the tests, they merge it into flox/staging, and when that passes more tests that's merged into flox/stable š | 23:32:45 |
| hexa (UTC+1) | so no relationship to release-25.05 whatsoever | 23:33:03 |
| hexa (UTC+1) | so stable users are fckd | 23:33:13 |
| SomeoneSerge (back on matrix) | I asked them about building the release channels, didn't get a positive reaction yet | 23:33:26 |
| SomeoneSerge (back on matrix) | I'll go ping one more time | 23:33:47 |
| hexa (UTC+1) | zowoq restored the cuda-stable jobset until 25.05 eol | 23:33:49 |
| hexa (UTC+1) | so there is some time | 23:33:58 |
| hexa (UTC+1) | but boy, do I not like the new deal | 23:34:05 |
| SomeoneSerge (back on matrix) | We'll fix it | 23:36:21 |
| SomeoneSerge (back on matrix) | I heard it's all paracetomol-induced autism | 23:42:11 |
| 26 Sep 2025 |
 ysndr | In reply to @ss:someonex.net
It's a tad confusing, but uhh they follow nixos-unstable, and when it passes the tests, they merge it into flox/staging, and when that passes more tests that's merged into flox/stable š A tad confusing indeed.
These stabilities are not describing the stability of the software that is contained, but the rate at which the software changes:
flox/unstable is rebased (almost) daily to the current nixos-unstable.
Every two weeks, that flox/staging is reset to that days unstable, every first Saturday of a month flox/stable is reset to the last staging. I.e. flox stable lags behind nixos-unstable by about 6weeks. | 08:40:59 |
| ysndr | In reply to @ss:someonex.net
I asked them about building the release channels, didn't get a positive reaction yet Nope no release channels currently :/ | 08:43:14 |
