| 27 Jun 2026 |
 K900 | hexa can you ack https://github.com/NixOS/nixpkgs/pull/536047 | 14:22:58 |
| K900 | Cause we're going to hit that pretty soon I'm assuming | 14:23:12 |
 hexa | queued | 14:23:47 |
 Myria | Another one that may need restarting: https://hydra.nixos.org/build/333248196 | 14:25:09 |
 emily | is hasDistutilsCxxPatch dead code now? | 14:25:54 |
| K900 | Uhh, possibly? | 14:26:11 |
| emily | ah, not for NumPy 1 | 14:26:11 |
| K900 | Not sure if there are any remaining uses | 14:26:17 |
| emily | ++ optionals (hasDistutilsCxxPatch && pythonOlder "3.12") [ | 14:26:18 |
| emily | though the CPython derivation implies it's only used for ,= 3.11 | 14:26:27 |
| emily | * though the CPython derivation implies it's only used for ≤ 3.11 | 14:26:30 |
| K900 | OK so very close to deadness | 14:26:31 |
| emily | 3.11 is due for drop this release right? | 14:26:57 |
| Myria | Yeah, distutils is apparently deprecated for python 3.12 from what I say in the numpy code | 14:27:01 |
| Myria | No sorry, removed in 3.12 | 14:27:40 |
| K900 | From upstream, yes | 14:28:03 |
| K900 | Numpy vendored its own | 14:28:07 |
| K900 | For a very long time | 14:28:09 |
| K900 | Which is what's being patched here | 14:28:17 |
| Myria | Ah, fun fun | 14:28:26 |
| emily | yeah but if the patch in CPython itself is conditioned on ≤ 3.11 then presumably the patch in NumPy 1 could be too | 14:28:39 |
| emily | or hm | 14:28:43 |
| emily | I guess no because the whole point is that it's an old distutils? | 14:28:50 |
 Alyssa Ross | Restarted curl.x86_64-linux, which was failed because of an invalid store path | 17:33:07 |
| Alyssa Ross | Do I also need to restart dependents? | 17:34:45 |
| K900 | python-jq broke | 17:36:45 |
| K900 | (again) | 17:36:50 |
| K900 | Because it relies on exact jq output (again) | 17:36:59 |
| hexa | supported until 2027-10 | 17:38:04 |
| hexa | so drop after nixos 26.11 is branched | 17:38:17 |
