| 15 May 2024 |
 jakehamilton | In reply to @fomosapien:matrix.org
https://go.theregister.com/feed/www.theregister.com/2024/05/14/nix_forked_but_over_politics/
I was expecting to see Lix xD Happy to take the hit so Lix can just keep doing good stuff quietly lol | 03:32:01 |
 toastal-matrix-sucks | nix (Lix, like Nix) 2.90.0-beta.1-lixpre20240506-b6799ab got it | 04:26:48 |
| toastal-matrix-sucks | FAQs says Flakes stabilization is a priority. I saw an open issue about unifying Nixpkgs fetchers & Flake fetchers as there is a lot of re-implementation of fetching & Flakes can’t inputs can’t accept (by not having the API for) fetching from as many sources … or mirrors for that matter. Is this a part of Lix’s goal plans too? | 04:30:20 |
| toastal-matrix-sucks | …And does Lix’s Flake need pre-commit-hook as an input that everyone downstream has to fetch? | 04:34:47 |
| toastal-matrix-sucks | (can there be a devInputs??) | 04:35:11 |
| toastal-matrix-sucks | …Or is the idea to not prefer Flakes, but just make sure Flakes are stable? | 04:50:50 |
| toastal-matrix-sucks | * …Or is the idea to not prefer or improve Flakes, but just make sure Flakes are stable & usable? | 05:03:59 |
| toastal-matrix-sucks | * …Or is the idea to not prefer or improve Flakes, but just make sure Flakes are stable & usable? FAQs wasn’t clear enough to me what the goals are. | 05:05:25 |
 Qyriad | Flake stablization is not a priority; the FAQ talks about compatibility | 05:06:40 |
| Qyriad | We want to:
make flakes better make Nix without flakes better decouple flakes and their features from each other and from Lix itself some kind of re-thought design of the problems flakes attempt to address (whether that is a flakes v2, something entirely new, or anything else)
| 05:09:04 |
| Qyriad | We want to:
make flakes better make Nix without flakes better decouple flakes and their features from each other and from Lix itself some kind of re-thought design of the problems flakes attempt to address (whether that is a flakes v2, something entirely new, or anything else)
| 05:09:16 |
| Qyriad | None of this is immediate priority, which is largely making Lix more stable, bug-free, and usable; but we are all thinking about it and occsionally brainstorming | 05:10:57 |
| toastal-matrix-sucks | Flakes v2 would be nice. I would use the alternatives, but they aren’t any better in my experience. | 05:14:46 |
| toastal-matrix-sucks | Thanks for the feedback | 05:14:54 |
 julia | so remember how I mentioned the other day that nix develop doesn't work when there are spaces in the path? I can do one better: it can't even copy a folder with spaces to the nix store... | 05:18:14 |
| julia | 
Download image.png | 05:18:15 |
| julia | * so remember how I mentioned the other day that nix develop doesn't work when there are spaces in the path? I can do one better: it can't even copy a folder with spaces to the nix store... ran into this with building a package based on the directory | 05:20:04 |
| Qyriad | That one is not surprising, because path literals set the store path name part to the directory name, and top-level store paths cannot contain spaces. But please gods tell me that "${builtins.path { path = ./.; name = "without-spaces"; }" works 😨 | 05:20:38 |
| julia | thankfully, yes. | 05:21:25 |
| Qyriad | Okay phew | 05:21:42 |
| julia | now to add it to this project lol | 05:21:46 |
| julia | I love workarounds | 05:21:56 |
| Qyriad | It would probably be nice though if we automatically converted top-level path names that contain spaces… I feel like there are probably a dozen lurking soundness errors in any attempt to do that | 05:22:24 |
| julia | I mean at least for anything involving like, mkDerivation you already have a name, so like, I feel like grabbing src that way would at least help... thou doesn't help with anything like filtersource etc | 05:23:18 |
| Qyriad | The (relatively) new Nixpkgs fileset API might not have this problem. Not sure | 05:23:51 |
| julia | i was using pkgs.nix-gitignore.gitignoreSource tbh | 05:24:33 |
| Qyriad | Ah | 05:24:37 |
| julia | lemme see, though... | 05:25:00 |
| julia | yeah it does fix it, by making the name always "source" | 05:26:13 |
| Qyriad | Ah yes because we need even more store paths with non-descriptive names, oh boy. But at least it does not have the space sproblem | 05:27:15 |
