Nix Documentation | 419 Members | |
| Discussion about documentation improvements around the Nix ecosystem | 86 Servers |
| Sender | Message | Time |
|---|---|---|
| 23 Sep 2024 | ||
In reply to @phanirithvij:matrix.orgin principle, but we have to be careful to rate limit that very tightly. it shouldn't run on every hydra job but more like once a month at most | 20:46:56 | |
| and it shouldn't be a blocker for anything | 20:47:29 | |
| What we could and should do is run linkchecks on all of our domains: nixos.org, nix.dev, ... | 20:48:28 | |
| * What we could and should do is unconditionally run linkchecks on all of our domains: nixos.org, nix.dev, ... | 20:48:35 | |
| * What we could and should do is more thoroughly run linkchecks on all of our domains: nixos.org, nix.dev, ... | 20:48:43 | |
| (but probably also not every single time) | 20:49:09 | |
| yes, scheduled github action (once/twice a month) makes sense, along with a maintainer script for running manually. | 20:55:11 | |
In reply to @fricklerhandwerk:matrix.orgnix.dev (incl nix manual), nixos-homepage, nix (+manual), nixpkgs (nixos, nixpkgs manuals), nixos-wiki-infra all seem to need the lychee checks. maybe need to think about more than simply adding sheduled gha to all these repos? | 21:00:23 | |
In reply to @fricklerhandwerk:matrix.org* nix.dev (incl nix manual), nixos-homepage, nix (+manual), nixpkgs (nixos, nixpkgs manuals), nixos-wiki-infra all seem to need the lychee checks. maybe need to think about it more than simply adding sheduled gha to all these repos? | 21:00:41 | |
In reply to @phanirithvij:matrix.orgSure. But adding checks is a fist step that's not wrong | 21:00:50 | |
In reply to @phanirithvij:matrix.org* Sure. But adding checks as a first step is not wrong, it won't make things worse | 21:01:09 | |
| (nix.dev has link checking and it regularly messes up builds because someone stops tolerating us hitting them) | 21:01:29 | |
| I don't think lychee check should be a blocker anywhere, it should just be a like warnings/report. | 21:03:53 | |
| * Sure. But adding checks as a first step is not wrong, it will prevent things from falling apart and gives us time to think | 21:03:56 | |
| * I don't think lychee check should be a blocker anywhere, it should just be like warnings/report. | 21:04:11 | |
| * Sure. But adding checks as a first step is not wrong, it will prevent things from falling apart while we're thinking | 21:04:21 | |
In reply to @fricklerhandwerk:matrix.orgok, linkcheck from sphinx. my experience with lychee with github urls even with github_token provided will still result in flaky 403 errors, I wonder how linkcheck is working fine | 21:11:42 | |
also the redirect are good to know I guess, they can also be fixed. | 21:12:10 | |
In reply to @fricklerhandwerk:matrix.orgdo you mean checks like this sphinx/linkcheck? | 21:13:04 | |
| does it work outside of sphinx? | 21:13:27 | |
In reply to @phanirithvij:matrix.orgNo, nix.dev linkchecks are pure Sphinix | 21:14:58 | |
In reply to @phanirithvij:matrix.org* No, nix.dev linkchecks are pure Sphinx | 21:15:15 | |
| I think scheduled gha is the way to go for external links | 21:15:30 | |
| am I right in assuming nix+nixpkgs+nixos manuals and wiki.nixos.org have a lot of overlap in people contributing? | 21:16:55 | |
In reply to @phanirithvij:matrix.orgEh, not that much. There are like 2-3 regular contributors who have touched all of them, but otherwise these are mostly disjoint groups as far as I can tell | 21:19:06 | |
| maybe a single gha which checks external links among all manuals + home-pages + wiki-dump could work? | 21:19:23 | |
| wait, need to think about rate-limits as well. | 21:19:32 | |
In reply to @phanirithvij:matrix.org* Eh, not that much. There are like 2-3 regular contributors who have touched all of them, but otherwise these are mostly disjoint groups as far as I can tell. I mean, it's complicated. Define "a lot". There's not that many people working on documentation overall. | 21:20:30 | |
| * Eh, not that much. There are like 2-3 regular contributors who have touched all of them, but otherwise these are mostly disjoint groups as far as I can tell. I may be wrong. I mean, it's complicated. Define "a lot". There's not that many people working on documentation overall. | 21:20:53 | |
| I found wiki to be easier to contribute to definitely | 21:21:24 | |