| 15 Mar 2024 |
 ryantm | Maybe I can trick them by doing something like thing = callPackage ./path/to/package/ { inherit stdenvNoCC;} | 23:12:01 |
 infinisil | Wouldn't work, it's too smart for that :P | 23:12:54 |
| infinisil | ryantm: How about not making it a top-level package? I can see nix-build -A nixpkgsManual.<something> | 23:13:40 |
| ryantm | That's a good idea. | 23:14:15 |
| infinisil | You can cheat the check by doing inherit ({ foo = callPackage ...; }) foo;, but that's not meant for this kind of thing | 23:14:18 |
| infinisil | ryantm: Why do you need it to be exposed via the top-level attributes anyways? | 23:15:32 |
| ryantm | infinisil: I just though it'd be nice to run nix run .#nixpkgs-manual-mmdoc-watch probably not required. | 23:16:55 |
| infinisil | ryantm: I feel like exposing a separate attribute in flake.nix would be better | 23:18:30 |
| infinisil | ryantm: Like, define it in doc/default.nix primarily, but then do nixpkgsManual = import ./doc or so in flake.nix | 23:19:04 |
| infinisil | (also camelCase is the convention in Nixpkgs, not snake_case) | 23:19:52 |
| infinisil | But I also get the argument for wanting the same interface between flakes and non-flakes, which wouldn't work with this.. | 23:21:55 |
