What do you mean? This is for CI, whatever the dev machine runs is irrelevant. Either way, the CI uses OCI containers built with Nix, you don't need Nix installed on the host. Also Nix supports MacOS.
I wasn't talking about developers machines. If you want to build and test software on Windows or Mac in CI, you obviously can't use a CI system that requires Linux. Yes I know about cross-compilation.
It is containers. It's based on Nixery, which is a virtual Docker registry where it puts together containers for you on the fly containing whatever packages you want from Nixpkgs.
By containers I mean I get to define my CI environment with containers, not nix or that somewhere under the hood containers are being used
Like Argo or Jenkins. Pushing nix as the DX for GHA equivalent was a poor choice by Tangled IMO. It's too unusual for your average dev, I'm not interested in learning nix so I can use CI.
Your username is vaguely familiar. You're a CUE enthusiast, right? I think I've read (and enjoyed and benefited from) some documentation you've written!
Yeah, I think it makes sense to also let people point to arbitrary OCI registries. I'd bet support for that is coming, especially since the execution environment is Dockerized anyway.
> Pushing nix as the DX for GHA equivalent
I think something like Nix actually makes more sense than YAML for this kind of thing. You want a DSL that is purpose-built so that configurations are declarative, simple configurations are simple to write, and configurations are composable. YAML is too much of a straightjacket. Some kind of built-in support for deep merges is a must, imo.
How powerful/expressive the language should be is debatable, I think. I'm interested in Turing-complete DSLs like Nix and Nickel, but CUE could be a good fit here, too.
Anyway I'm sure they'll add first-class support for using some OCI artifact to define a CI environment. Looks like their CI implementation only recently entered the first alpha.
