What are your plans for NixOS 23.05?

1

With the NixOS 22.11 :raccoon: release out the door, I fondly remember the inspiring threads where people announced their big plans for the upcoming release cycle.

Those threads can be a good opportunity to find people with similar interests who may be interested in some form of collaboration.

Feel free to start and post yours, while I contemplate mine. :laughing:

5 Likes
2

Nuking nodePackages now that buildNpmPackage exists.

3 Likes
3

Having linux binaries from the Internet to work natively :star_struck:
It’s really a show stopper NixOS adoption for many :confused:

3 Likes
4
  1. upstream as many local pkgs and modules as humanly possible (I want to carry 0 patches to nixpkgs)
  2. flakify openhab
  3. be better at helping out new people
3 Likes
5

Reach out if you’d like to assist in planning (I have a rough timeline in my head that hinges on some fixes landing) :sparkles:

6

Wait, does that mean submitting it into nixpkgs or just some downstream repo? Would be nice to have some competition to home-assistant in nixpkgs!

That being said I’d like to finally get support for custom components and lovelace modules in home-assistant into shape. It has been sitting too long already.

https://github.com/NixOS/nixpkgs/pull/160346

1 Like
7

I have thousands of lines of modules and helpers for handling my openhab setup which I want to get into nixpkgs proper (openhab is lovely and so much more pleasant than HA in my experience), but there are a few things that need to happen first, so the plan is:

  1. clear separation of upstreamable configuration, upstreamable helpers, personal configuration and personal helpers (it’s actually quite cleanly separated but I need to make sure it’s 100%)
  2. expose it as a flake
  3. get input from others who are interested in using it as well in order to make it ready for upstreaming
  4. finally upstream it into nixpkgs

Because it’s so opinionated and there is so much “code”, it’s unlikely that it can be upstreamed in its current form and then having to deal with making my own stuff fit whatever is considered acceptable for upstreaming, is just going to be a ton of very unpleasant work.

Making it available for testing as a flake at least allows us (me and whoever else is interested) to iterate much faster before upstreaming into nixpkgs proper.

The next question is obviously: when?

Step 1 is ongoing. I’m hoping to find some time over xmas to get that done, but realistically it’s going to take a couple of months. I do consider it more important to have the flakified module made available to the public than having it merged into nixpkgs because getting more eyes on it will simply give the biggest payback.

Hopefully all done by xmas next year.

EDIT: details on the thousands comment:

$ wc -l < modules/toupstream/services/openhab/*
1855

This excludes personal helpers for which a number probably should be upstreamed too.

5 Likes