Robert Hensing

Here's a possible implementation - https://github.com/hercules-ci/hercules-ci-agent/pull/516 The flaws are - Compensating for bad subflake support in Nix, while the subflake is actually a workaround for https://github.com/NixOS/nix/issues/7730 - Re-evaluating the modules...

:tada: You may now use [flake-parts.flakeModules.partitions](https://flake.parts/options/flake-parts-partitions) to achieve this. ----- I'm debating whether to add an opinionated module that sets ```nix partitionedAttrs.checks = "dev"; partitionedAttrs.devShells = "dev"; ``` and perhaps...

I think the way to go here is to make the website repo expose a function for this purpose. The documentation generation story is in flux at the moment, with...

I'm quite happy to accept PRs to https://github.com/hercules-ci/flake.parts-website, after which you can use that repo to build your own docs. Doing it this way also helps with discoverability of your...

If your module is already published, you can do something like the following. I'll use `hercules-ci-effects` as an example ``` cd hercules-ci-effects nix build github:hercules-ci/flake.parts-website --override-input hercules-ci-effects . ``` You...

Arguably the `nixos*` options should move there as well, although those have the bad excuse of `nix flake check` comitting the same layering violation.

Flake evaluation is designed not to depend on external factors, such as which host is evaluating the flake. I would recommend something like NixOS or home-manager to configure host environments....

Unlike `NixOS`, the attributes of a flake are often accessed individually, which means that laziness plays a more important role; somewhat comparable to Nixpkgs. Assertions require more to be evaluated,...

Good idea. The root `nixpkgs.*` options could be permanent aliases for `nixpkgsSets.pkgs` or something.

For Rosetta, you can already get a `pkgs` via `getSystem` and/or `withSystem`. Those will also give you access to your own `packages` for that system. The feature still makes a...