| 19 Jan 2025 |
 WeetHet | Wait it does? | 15:08:44 |
 Tristan Ross | Yes | 15:08:51 |
 emily | what would you do to debug/fix it that wouldn't change the original derivation anyway? | 15:08:57 |
| Tristan Ross | env-vars is the file | 15:09:10 |
| WeetHet | Skipping tests that require network/flaky tests | 15:09:33 |
| WeetHet | To be fair, it's probably a non-issue when you have something like crate2nix | 15:10:33 |
| WeetHet | * To be fair, it's probably a non-issue when you have something like crate2nix, since rebuilds don't take that much | 15:10:44 |
| emily | fair enough | 15:11:04 |
| WeetHet | I still wish there was a way to make each phase it's own derivation without needing to copy the build folder ten times | 15:11:47 |
| WeetHet | * I still wish there was a way to make each phase its own derivation without needing to copy the build folder ten times | 15:11:53 |
| emily | yes, it would be good. it requires much more fundamental changes than Nixpkgs can do though | 15:13:04 |
| WeetHet | On the other hand, again, copying isn't such a big deal with modern CoW filesystems | 15:14:42 |
| emily | it would go from not in the store to in the store, so usually you wouldn't be able to CoW | 15:19:19 |
| emily | (even when you have /nix/tmp and /nix/store on the same backing partition, there are technical reasons related to the bind mount stuff why it doesn't actually work on Linux) | 15:19:43 |
| 22 Jan 2025 |
|  Gaétan Lepage joined the room. | 11:15:29 |
