!9IQChSjwSHXPPWTa:lix.systems

Lix

807 Members
Lix user channel. Feel free to discuss on-topic issues here and give each other help. For matrix.to links to the rest of the Lix channels, see: https://wiki.lix.systems/books/lix-organisation/page/matrix-rooms235 Servers

Load older messages


SenderMessageTime
21 Feb 2025
@lunaphied:lunaphied.meLunaphiedAh18:51:27
@aidalgol:matrix.orgaidalgolhttps://discourse.nixos.org/t/lix-evaluates-my-rust-crane-package-instantly-nix-2-24-takes-2-mins19:46:11
@aidalgol:matrix.orgaidalgolhaha! oh no...19:46:25
@aidalgol:matrix.orgaidalgol* haha! oh no... 🫠19:46:29
@raitobezarius:matrix.orgraitobezariusWhat20:31:03
@raitobezarius:matrix.orgraitobezariusThey broke IFD performance that hard20:37:52
@raitobezarius:matrix.orgraitobezariusImpressive20:37:58
@chloe:angelwood.xyz@chloe:angelwood.xyz removed their display name Chloe.21:10:05
@chloe:angelwood.xyz@chloe:angelwood.xyz 21:10:16
@chloe:angelwood.xyz@chloe:angelwood.xyz 21:11:25
@chloe:angelwood.xyz@chloe:angelwood.xyz 21:12:29
@chloe:angelwood.xyz@chloe:angelwood.xyz 21:12:37
@lunaphied:lunaphied.meLunaphied Sounds like it's just eval cache brokenness 21:33:21
@just1602:systemli.orgjust1602I probably don't understand the complexity of such setup, but I don't think I understand why there's no tests to catch this kind of things.21:37:13
@just1602:systemli.orgjust1602Weither it's benchmark / stresstest, some sort of property testing. I mean, there must be ways to properly test that, tho?21:37:50
@delroth:delroth.netdelrothif evals are too fast then detsys can't sell you flakehub "pre-resolved paths" /s21:45:55
22 Feb 2025
@kfears:matrix.orgKFears (no longer human)
In reply to @whovian9369:matrix.org

If that was for me, the only reference that I could find for the rev string was in my flake.lock for /lix-project/nixos-module

$  rg -N --context=4 621aae0f3cceaffa6d73a4fb0f89c08d338d729e
flake.lock
      "locked": {
        "lastModified": 1738176840,
        "narHash": "sha256-NG3IRvRs3u3btVCN861FqHvgOwqcNT/Oy6PBG86F5/E=",
        "ref": "refs/heads/main",
        "rev": "621aae0f3cceaffa6d73a4fb0f89c08d338d729e",
        "revCount": 133,
        "type": "git",
        "url": "https://git.lix.systems/lix-project/nixos-module"
      },
Hmm, could you check git branch and git tag -a?
01:03:03
@kfears:matrix.orgKFears (no longer human) Another a little sus thing is ref being refs/heads/main. Can you show how you added the entry in flake.nix? Perhaps that might shed some light on the issue 01:04:39
@kfears:matrix.orgKFears (no longer human) As for why CppNix doesn't throw a warning, it's because Lix is using git CLI and CppNix is using libgit2 the C library, I think. I doubt that libgit2 prints any warnings to the terminal anyway, and there are some behavior differences anyway 01:07:31
@kfears:matrix.orgKFears (no longer human)I suspect that CppNix would be just as confused, just silently. But I think the warning is useful, and it would be nice to investigate01:08:08
@whovian9369:matrix.orgWhovian9369 Onto what, a clone of nixos-module? 01:08:30
@whovian9369:matrix.orgWhovian9369
lix = {
  url = "git+https://git@git.lix.systems/lix-project/lix";
    /*
      Future me, the pattern for using Forgejo URLs is:
      git+https://git@${domain}/${user_org}/${repo}?ref=refs/tags/${TAG}
      git+https://git@${domain}/${user_org}/${repo}?rev=${commitHash}
    */
  inputs.nixpkgs.follows = "nixpkgs";
  inputs.flake-compat.follows = "flake-compat";
};

lix-module = {
  url = "git+https://git.lix.systems/lix-project/nixos-module";
    /*
      Future me, the pattern for using Forgejo URLs is:
      git+https://git@${domain}/${user_org}/${repo}?ref=refs/tags/${TAG}
      git+https://git@${domain}/${user_org}/${repo}?rev=${commitHash}
    */
  inputs.lix.follows = "lix";
  inputs.nixpkgs.follows = "nixpkgs";
  inputs.flake-utils.follows = "flake-utils";
};
01:09:03
@kfears:matrix.orgKFears (no longer human) Hmm, interesting. The entry I'm using is url = "https://git.lix.systems/lix-project/nixos-module/archive/main.tar.gz. Can you check if it works? narHash in flake.lock shouldn't change, unless you haven't updated for a while 01:14:42
@kfears:matrix.orgKFears (no longer human) Oh, I think you'll have to update specifically the module input before checking narHash, because it seems you're missing 1 commit
https://git.lix.systems/lix-project/nixos-module/commit/8aaf8cff6e2b8c7d19abf7d73f4d04197223a480
01:18:23
@whovian9369:matrix.orgWhovian9369 Would I want to nix flake update before or after I swap out the URL? :P 01:18:52
@kfears:matrix.orgKFears (no longer human)Before. I'm not sure if it will reload automatically or you'll also have to run it after updating01:19:34
@whovian9369:matrix.orgWhovian9369

I'm not sure if it will reload automatically
I bet nix flake update would yell since it doesn't match.

01:20:18
@whovian9369:matrix.orgWhovian9369 *

I'm not sure if it will reload automatically
I bet nix flake update would yell since it doesn't match.

01:20:23
@whovian9369:matrix.orgWhovian9369 *

I'm not sure if it will reload automatically

I bet nix flake update would yell since it doesn't match.

01:20:28
@whovian9369:matrix.orgWhovian9369Well it'd update the hash, I mean.01:20:39

Show newer messages


Back to Room ListRoom Version: 10