1. 17
  1.  

    1. 25

      Declaring flakes stable and merging Eelco’s draft PRs in their fork sounds like giving up on fixing the CppNix leadership.
      This isn’t really a surprise given they seem more focused on the “determinate community” than the Nix one, and they had previously announced wanting to provide flake stability in their fork/distribution. It’s nonetheless disappointing to see play out.

      1. 9

        It’s sad, but it’s true that the flake stuff is de-facto stable, it’s here to stay, there are too many people using it. Declaring it stable is primarily just recognizing this. I’m not defending them, but they made an effort to open the eyes of the community on just that some months ago, but of course that thread ended in nothing, and in nix we are still with experimental flakes, that sucks for true

        1. 6

          Yeah I use flakes and am not against trying to stabilize some of it! I’d also like to see alternatives actually explored, though at this point they’d need a bridge with flakes to be viable (flake-compat at least proves it’s possible).

          The Nix project has had a leadership issue for a while and that’s what’s going to hurt Nix more in the long term IMO, at least CppNix. I hope that eventually Nixpkgs can endorse a Nix version that has functional leadership. That wouldn’t require any Determinate involvement and would essentially carry over most of the community.

          1. 3

            The thing I do not understand is that… They can merge. Whenever they want. This is Eelco project. DetSys is Eelco’s company.

            What is the problem that make him merge it in private, not in public?

            1. 7

              No we can’t, no he can’t, not anymore. He’s one of five of the Nix team that operates on a consensus basis.

              1. -1

                That… Is cute but he actually could. Anyway.

        2. 8

          It looks like the major thing this does is create some enterprise support for flakes (in their current form).

          There’s still a lot of fertile ground for improving flakes, and I bet that upstream Nix will keep that up, but for commercial use having people that can be paid money to guarantee behavior is pretty helpful. Overall, another step forward in getting Nix more palatable to industry. Good work.

          1. 6

            Regardless of the governance politics, is determinate nix useful or pleasant compared to a plain old flake? Are people using this and flakehub much?

            1. 3

              I have not used Determinate Nix, but I think you can infer their main selling points from their marketing copy: they provide some minor UX and DX improvements alongside a lot of enterprise integrations and support guarantees. Upstream Nix is not focused on catering to things like SOC2 compliance or MDM integration, so it makes sense for a private company to focus their efforts there.

              1. 2

                Looking for the things ai wanted, seems like it still only supports paths, tarballs, Git, & Mercurial—with no support of having mirrors to fallback to for when servers inevitably go down. I was hoping these would have been address despite Nixpkgs fetchers support going beyond these two limitiations.

                1. 1

                  I don’t know about Linux, but it’s the best option on macOS. The uninstaller and ability to persist across major OS upgrades are really nice.

                2. 1

                  Can we use determinate Nix without nixd? My reason is that I assume things like parallel evaluation will be implemented in Nix and not nixd. I have zero interest on nixd but I would really appreciate parallel evaluation.

                  1. 2

                    Sure, you can get the source from https://github.com/DeterminateSystems/nix-src. It is fully functional on its own. Obviously we don’t support using it in that fashion, but you probably didn’t care about that anyway.