Nix Flakes | 859 Members | |
| 177 Servers |
| Sender | Message | Time |
|---|---|---|
| 18 Mar 2025 | ||
| only a subset of things block the channel from advancing | 15:49:04 | |
if all failed builds stopped unstable from bumping, it'd never bump | 15:49:12 | |
| but we do have e.g. automated end-to-end testing of installations, common graphical environments, etc. | 15:49:24 | |
| Yeah it does elikoga, that's what I expected. Being a bit difficult to test whether it launches or not | 15:49:28 | |
| (since we have ~100k builds and many of them are always broken 🫠) | 15:50:18 | |
| I wish it was a little bit clearer when a package is stable or not? | 15:50:37 | |
| Would it be possible to add something to the nixpkgs search interface to link or show like a red/yellow/green bubble per platform about how "stable" it is? Whether builds have errors etc? | 15:52:40 | |
| That would be useful | 15:52:46 | |
| Or even just a link to the relevant page in hydra would already be helpful | 15:53:18 | |
| 19 Mar 2025 | ||
| 16:17:32 | ||
| 20 Mar 2025 | ||
| 03:15:29 | ||
| 03:19:04 | ||
In reply to @laauurraa:matrix.orgNobody built it yet, so it's not there | 11:39:34 | |
| * Nobody built that integration yet, so it's not there | 11:39:47 | |
| I don't believe that it's trivial to link the different projects without pulling your hair out in frustration | 11:41:06 | |
| 16:20:06 | ||
| I'm having issues with github auth when updating my flake [jono@dobro:~/sync/configs/nix]$ nix flake update --flake .
(base) I get a 200. | 16:20:31 | |
| * I'm having issues with github auth when updating my flake
if I directly curl that https://api.github.com/repos/nix-community/disko/commits/HEAD I get a 200. | 16:21:05 | |
| 21 Mar 2025 | ||
TIL: outputs = import ./function.nix; is not allowed | 11:38:19 | |
| wonder if that's a bug | 11:38:24 | |
| In flake.nix? That's expected, it's not a real Nix file | 11:39:00 | |
| There's quite a bit of restrictions on the flake.nix format that people complain about sometimes | 11:39:06 | |
for inputs I expect that by now, yes | 11:39:10 | |
for outputs I don't think there is any rational point | 11:39:16 | |
| and the error seems like it's probably just an implementation bug | 11:39:24 | |
| ("expected a function but got a thunk") | 11:39:31 | |
| No that sounds like the format restriction+bad error messages | 11:39:46 | |
oh, I guess it's because outputs's arguments get introspected to define inputs 🙃 | 11:39:47 | |
| when you don't explicitly list them | 11:39:56 | |
| (and it uses the registry inputs, which are pinned by default in some setups, so pretty footgunny) | 11:40:11 | |