7 ms·

The reception on the nixos discourse is worth a look: https://discourse.nixos.org/t/introducing-flakehub/32044 https://discourse.nixos.org/t/introducing-flakehu

The reception on the nixos discourse is worth a look: https://discourse.nixos.org/t/introducing-flakehub/32044 https://discourse.nixos.org/t/introducing-flakehub/32044

MuffinFlavored 3y ago
https://discourse.nixos.org/t/introducing-flakehub/32044/3 https://discourse.nixos.org/t/introducing-flakehub/32044/3 I am very disappointed. Instead of focusing on actually fixing problems with Flakes and trying to stabilize them, DetSys goes full in on experimental Flakes, advertising them as the future, ignoring all problems with them and building products on top of them. Meanwhile everybody trying to help out official efforts, the Nix team, Nixpkgs maintainers, the documentation team and more are suffering because Flakes isn’t stable. It seems like DetSys has no interest in alleviating the pressure and bridging the gap that Flakes has created in the community. And more concretely, it’s a well-known problem that Flakes suffer from an explosion of dependencies 64 with little reuse and interoperability. The design of how dependencies are locked just isn’t right. But oh surprise, that’s exactly what FlakeHub seems to help with by adding an ad-hoc third-party versioning scheme on top of Flakes. And all that centralised (wasn’t the whole point of Flakes to not be centralised?), proprietary and with seemingly no intention of upstreaming it in any way. This all just doesn’t sit right with me :frowning: I don't want to discredit what this person is saying at all but does anybody else read things like this in regards to open source software and instantly think "sheesh, that person could benefit from going outside and touching some grass every once in a while?" Oh, and it has 50 likes, lol
- pxc 3y ago
  Infinisil has over 600 commits in the Nixpkgs repo alone, and is also active in other codebases in the ecosystem. This comment on Discourse has a lot of likes in part because people in the community recognize his username from his contributions even if they haven't worked with him. He's not some random passerby throwing out an opinion about a project he has no stake in. He's a longtime, highly active contributor in the Nix ecosystem who cares about projects he's involved in and other efforts related to them. You don't have to agree with his take but this is disrespectful and lame.
  - r3trohack3r 3y ago
    As an outsider I’m super confused reading this discourse. There are a few people I consider O.G. Nix hackers in there. edolstra isn’t some rando either, he’s behind (if I understand correctly) the original Nix paper all the way back in 2004: https://edolstra.github.io/pubs/nspfssd-lisa2004-final.pdf https://edolstra.github.io/pubs/nspfssd-lisa2004-final.pdf
    - pxc 3y ago
      Yeah, Eelco is the original author of Nix and continues to to maintain it. I think all you're missing is my point. ;) My point is not that it's automatically a huge deal whether Infinisil (or anyone) disapproves of this or that development in the community. But that nobody deserves to be dismissed with a backhanded I-hope-my-attempt-to-discredit-you-doesn't-discredit-you 'go touch grass' comment. And that I think within the Nix(OS) community there's basic neighborly recognition that is clearly missing here (on HN).
      - r3trohack3r 3y ago
        Ah! I misunderstood your comment. I thought you meant the original post(s) on discourse were disrespectful. Totally agree that “go touch grass” is a disrespectful take. These are a bunch of O.G. hackers who’ve invested a portion of their life into a significantly underfunded problem statement in our industry. They deserve respect and our industry is better for them spending their time the way they did. I’m personally very thankful these folks exist and are as passionate as they are. It gives me hope that one day I’ll work with a package manager in production that does more than just YOLO unpack an archive file into a global namespace followed by some hopefully deterministic shell scripts.
  - MuffinFlavored 3y ago
    > Infinisil has over 600 commits in the Nixpkgs repo alone, and is also active in other codebases in the ecosystem. > people in the community recognize his username from his contributions > He's a longtime, highly active contributor in the Nix ecosystem who cares about projects he's involved in and other efforts related to them. > You don't have to agree with his take but this is disrespectful and lame. Without doubling down on my viewpoint that you already think is lame and disrespectful, you don't think any of these achievements are probably mildly unhealthy/obsession level?
grhmc 3y ago
I just posted a reply on Discourse (https://discourse.nixos.org/t/introducing-flakehub/32044/14?u=grahamc https://discourse.nixos.org/t/introducing-flakehub/32044/14?...) but for posterity: I’m sorry some folks aren’t excited about FlakeHub, but that is okay. We’ve made it because it solves real problems we’ve experienced, and I think it is going to make a big difference in the community. We’re all contributing to the project in the ways we want to, and this is an example of that. We see flakes as the future of Nix, and we’re unabashedly working to make that true. One way we’re doing that is using flakes extensively every day and exposing ourselves to the bugs, papercuts, and pain points. We often deeply research and resolve these problems upstream, without fanfare or posting on Discourse. I think we have the common goal of growing and improving Nix, and we’re glad to be doing it together. We would love to see a continuation of RFC-144 accepted into Nix upstream. The version resolving feature of FlakeHub is a bandaid that works until the upstream project has a comparable feature. In my opinion, it is a good thing to do experiments like this without committing the entire project to it.
- zamalek 3y ago
  As flawed as flakes are, they are significantly better than channels and all the hacks that surround them. Flakes are how I manage my system today and I honestly believe that there is no harm in building tooling around them - we all know that things are subject to change. This is now in my configs: https://codeberg.org/jcdickinson/nix/src/branch/main/flake.nix https://codeberg.org/jcdickinson/nix/src/branch/main/flake.n... Something I've noticed: `nixos/*` and `home-manager/*` are incompatible (home-manager expects 23.11 while nixos is 23.05).
  - grhmc 3y ago
    Very cool! You can get nixpkgs unstable like this: nixpkgs-unstable.url = "https://flakehub.com/f/NixOS/nixpkgs/1.*.tar.gz";

The reception on the nixos discourse is worth a look: https://discourse.nixos.org/t/introducing-flakehub/32044 https://discourse.nixos.org/t/introducing-flakehub/32044

MuffinFlavored 3y ago

https://discourse.nixos.org/t/introducing-flakehub/32044/3 https://discourse.nixos.org/t/introducing-flakehub/32044/3 I am very disappointed. Instead of focusing on actually fixing problems with Flakes and trying to stabilize them, DetSys goes full in on experimental Flakes, advertising them as the future, ignoring all problems with them and building products on top of them. Meanwhile everybody trying to help out official efforts, the Nix team, Nixpkgs maintainers, the documentation team and more are suffering because Flakes isn’t stable. It seems like DetSys has no interest in alleviating the pressure and bridging the gap that Flakes has created in the community. And more concretely, it’s a well-known problem that Flakes suffer from an explosion of dependencies 64 with little reuse and interoperability. The design of how dependencies are locked just isn’t right. But oh surprise, that’s exactly what FlakeHub seems to help with by adding an ad-hoc third-party versioning scheme on top of Flakes. And all that centralised (wasn’t the whole point of Flakes to not be centralised?), proprietary and with seemingly no intention of upstreaming it in any way. This all just doesn’t sit right with me :frowning: I don't want to discredit what this person is saying at all but does anybody else read things like this in regards to open source software and instantly think "sheesh, that person could benefit from going outside and touching some grass every once in a while?" Oh, and it has 50 likes, lol

pxc 3y ago
Infinisil has over 600 commits in the Nixpkgs repo alone, and is also active in other codebases in the ecosystem. This comment on Discourse has a lot of likes in part because people in the community recognize his username from his contributions even if they haven't worked with him. He's not some random passerby throwing out an opinion about a project he has no stake in. He's a longtime, highly active contributor in the Nix ecosystem who cares about projects he's involved in and other efforts related to them. You don't have to agree with his take but this is disrespectful and lame.
- r3trohack3r 3y ago
  As an outsider I’m super confused reading this discourse. There are a few people I consider O.G. Nix hackers in there. edolstra isn’t some rando either, he’s behind (if I understand correctly) the original Nix paper all the way back in 2004: https://edolstra.github.io/pubs/nspfssd-lisa2004-final.pdf https://edolstra.github.io/pubs/nspfssd-lisa2004-final.pdf
  - pxc 3y ago
    Yeah, Eelco is the original author of Nix and continues to to maintain it. I think all you're missing is my point. ;) My point is not that it's automatically a huge deal whether Infinisil (or anyone) disapproves of this or that development in the community. But that nobody deserves to be dismissed with a backhanded I-hope-my-attempt-to-discredit-you-doesn't-discredit-you 'go touch grass' comment. And that I think within the Nix(OS) community there's basic neighborly recognition that is clearly missing here (on HN).
    - r3trohack3r 3y ago
      Ah! I misunderstood your comment. I thought you meant the original post(s) on discourse were disrespectful. Totally agree that “go touch grass” is a disrespectful take. These are a bunch of O.G. hackers who’ve invested a portion of their life into a significantly underfunded problem statement in our industry. They deserve respect and our industry is better for them spending their time the way they did. I’m personally very thankful these folks exist and are as passionate as they are. It gives me hope that one day I’ll work with a package manager in production that does more than just YOLO unpack an archive file into a global namespace followed by some hopefully deterministic shell scripts.
- MuffinFlavored 3y ago
  > Infinisil has over 600 commits in the Nixpkgs repo alone, and is also active in other codebases in the ecosystem. > people in the community recognize his username from his contributions > He's a longtime, highly active contributor in the Nix ecosystem who cares about projects he's involved in and other efforts related to them. > You don't have to agree with his take but this is disrespectful and lame. Without doubling down on my viewpoint that you already think is lame and disrespectful, you don't think any of these achievements are probably mildly unhealthy/obsession level?

grhmc 3y ago

I just posted a reply on Discourse (https://discourse.nixos.org/t/introducing-flakehub/32044/14?u=grahamc https://discourse.nixos.org/t/introducing-flakehub/32044/14?...) but for posterity: I’m sorry some folks aren’t excited about FlakeHub, but that is okay. We’ve made it because it solves real problems we’ve experienced, and I think it is going to make a big difference in the community. We’re all contributing to the project in the ways we want to, and this is an example of that. We see flakes as the future of Nix, and we’re unabashedly working to make that true. One way we’re doing that is using flakes extensively every day and exposing ourselves to the bugs, papercuts, and pain points. We often deeply research and resolve these problems upstream, without fanfare or posting on Discourse. I think we have the common goal of growing and improving Nix, and we’re glad to be doing it together. We would love to see a continuation of RFC-144 accepted into Nix upstream. The version resolving feature of FlakeHub is a bandaid that works until the upstream project has a comparable feature. In my opinion, it is a good thing to do experiments like this without committing the entire project to it.

zamalek 3y ago
As flawed as flakes are, they are significantly better than channels and all the hacks that surround them. Flakes are how I manage my system today and I honestly believe that there is no harm in building tooling around them - we all know that things are subject to change. This is now in my configs: https://codeberg.org/jcdickinson/nix/src/branch/main/flake.nix https://codeberg.org/jcdickinson/nix/src/branch/main/flake.n... Something I've noticed: `nixos/*` and `home-manager/*` are incompatible (home-manager expects 23.11 while nixos is 23.05).
- grhmc 3y ago
  Very cool! You can get nixpkgs unstable like this: nixpkgs-unstable.url = "https://flakehub.com/f/NixOS/nixpkgs/1.*.tar.gz";