| 27 Sep 2021 |
 David Arnold (blaggacao) | Right hand side | 19:27:48 |
 @timdeh:matrix.org | ah | 19:28:01 |
| @timdeh:matrix.org | I don't know what you mean 😅 | 19:28:15 |
| David Arnold (blaggacao) | (meaning on the rhs of = in an output assignation) | 19:28:34 |
| @timdeh:matrix.org | ah okay lemme check | 19:28:47 |
| David Arnold (blaggacao) | If either self or inputs are somewhat used to infer the lhs then that usually ends up in an infinit recursion | 19:29:31 |
| David Arnold (blaggacao) | https://github.com/fluidattacks/makes/pull/575 | 19:30:34 |
| David Arnold (blaggacao) | I had that just recently | 19:30:40 |
| @timdeh:matrix.org | I guess it's back to the old "comment until the error stops" cycle 🤦♂️ | 19:30:50 |
| @timdeh:matrix.org | this seems to be happening at a more fundamental level, must be something during parsing of lock file because:
A. it happens almost instantly before any meaningful evaluation could possibly happen
B. I've commented out basically everything in my flake.nix and it's still happening | 19:36:08 |
| @timdeh:matrix.org | making it as minimal as possible:
digga.lib.mkFlake
{
inherit self;
inputs = {};
| 19:38:55 |
| @timdeh:matrix.org | this still causes the issue ❗️ | 19:39:02 |
| @timdeh:matrix.org | * making it as minimal as possible:
digga.lib.mkFlake
{
inherit self;
inputs = {};
}
| 19:39:21 |
| @timdeh:matrix.org | did the processing of self change recently? | 19:39:37 |
| @timdeh:matrix.org | nvm that can't be it, I just tried with an empty self | 19:40:06 |
| @timdeh:matrix.org | oh boy | 19:41:01 |
| @timdeh:matrix.org | no it's definitely mkFlake, as removing it and adding a minimal nixosConfigurations just using nixos.lib.nixosSystem doesn't cause it 🤔 | 19:45:46 |
| @timdeh:matrix.org | maybe it's something to do with my particular follows as I just messed with some to get rid of duplicates | 19:46:14 |
|  yusdacra changed their profile picture. | 19:50:41 |
| @timdeh:matrix.org | oh wow it is because of follows somehow, even using latest unstable nix. If I just comment out all my follows then it doesn't happen 🤯 | 20:00:57 |
| @timdeh:matrix.org | okay at least I found it, I guess digga was in a situation where it was following itself due to me not being careful 😅 | 20:17:27 |
| @timdeh:matrix.org | I think I'll open an upstream issue since that seems like something nix should be able to detect | 20:31:34 |
 Pacman99 | In reply to @blaggacao:matrix.org
That's not necessary, the submodule just checks out the respective commit. If that commit is already on GitHub's server then that's all it takes. Got it, just switched to the branch on devos in the submodule and ran bors try. It so far seems to be working now. | 22:25:54 |
| Pacman99 | In reply to @timdeh:matrix.org
oh wow it is because of follows somehow, even using latest unstable nix. If I just comment out all my follows then it doesn't happen 🤯 Out of curiosity, how did you manage to get digga to follow itself? | 22:26:33 |
| Pacman99 | That does seem like something nix should detect and report the recursion. | 22:27:46 |
| @timdeh:matrix.org | yeah it'd be better than just having a generic stack overflow and having to discover it yourself | 22:28:07 |
| @timdeh:matrix.org | so most of the flake-utils were following digga's flake-utils-plus/flake-utils input, so I was just carelessly copy pasting `*.follows = "digga/flake-utils-plus/flake-utils" and I happened to paste that on digga's flake-utils input itself 😅 | 22:28:55 |
| @timdeh:matrix.org | * so most of the flake-utils were following digga's flake-utils-plus/flake-utils input, so I was just carelessly copy pasting *.follows = "digga/flake-utils-plus/flake-utils" and I happened to paste that on digga's flake-utils input itself 😅 | 22:29:06 |
| Pacman99 | Ohh interesting yeah that makes sense | 22:35:12 |
| Pacman99 | Good to know for the future | 22:35:19 |
