Thanks for the link! When I get around to this I’ll message you about it.
I want systemd 242 in. This is what I’ll mostly be working on the coming month.
Things I’d like to make headway on:
- Specify an API for plugins/sub-packagesets https://github.com/NixOS/nixpkgs/issues/59344
- Start simplifying various parts of nixpkgs using said APIs
(would be happy to talk about the above on freenode/#nixos-wg-ai)
- Develop a homogeneous way to enable debugging for built packages (e.g. you can’t wrap R packages in
- Managing WINE with nix: https://github.com/deliciouslytyped/nix-winnix/
Things I’m ambiently thinking about:
- Integrating configurations managed by external applications with nix (home-manager)
- Generally standardizing things across nixpkgs
- UX improvements
There’s WIP on systemd 242: https://github.com/NixOS/nixpkgs/pull/61321
Another one that’s been discussed, but I haven’t had time to really look at:
- enable debugging symbols for all haskell packages
mentioned here: https://github.com/NixOS/nixpkgs/pull/52255
As already mentioned, I’m working more on updating static-haskell-nix with the intent of moving more or all of it into
nixpkgs, and getting Hydra CI it, because otherwise it can quickly break.
I also have a self-written module using lego; it’s a bit AWS Route53-specific, but the general idea is pretty extensible:
My wishlist would be:
- be able to join a domain with sssd (and be able to log-in with gdm)
- secure boot
For sssd, I played with it in the past. I think I was able to log-in with the console but not with gdm. I’m not sure I could pull it off but I might try again.
I’m pretty sure secure boot is beyond my capabilities. I wont even try.
It will be released in August 2019, so we can start migrate to beta in
Another one I just thought of, but don’t have time myself.
- Build a graphical installer for NixOS using the Calamares framework
If anyone would be interested in working on this, I think it would be greatly appreciated by new NixOS users!
I also hope that my PR for a significantly improved gstreamer (a lot more plugins building!) will make it in
Happy to report that this one was merged.
The official documentation is very limited. To be honest, it’s disappointing! It is quite difficult for new users to feel comfortable.
I really think that we can extend the
acme service to have enough options to support many acme clients out there such as lego or certbot itself. We have some kind of “interface” already for the http challenge (i.e. where other services expect certs to be available on disk or the preDelay script…), we “just” need to make sure we have an interface for the DNS challenge too
I’ve been fixing some bugs in the acme client the past month, which should be merged soon.
Working with it, the interface so far is pretty tightly coupled to both
nginx I think it would be better to just start a new module with the idea of multiple challenges baked in from the get-go. I would not try to make it acme-client agnostic though; that seems to add needless complexity. we keep the old
acme module and create a new
lego module that supports multiple challenges and what not. Changing the old
acme module to support multiple challenges and acme clients, whilst remaining backwards compatibility seems too much of a hassle to me.
This needs design first before implementation! It’s a big task to get it right. We need Usability/UX people here, no programmers!
That’s well known. See Documentation improvements for discussion and progress.
Another nice addition (which I don’t think I will have time for) would be version 10.1 of the CUDA toolkit. It seems that the packaging format for
.run files has changed (the Perl script is gone).
@danieldk Got some WIP for cudatoolkit 10.1 here: https://github.com/knedlsepp/nixpkgs/commit/630068ecdbe3f53901cee19ddf6a05675ce89766
It should build fine, but I didn’t unify the expression with the older releases.
Damn, duplicated efforts.
Got it unify here: https://github.com/dguibert/nixpkgs/commit/29933b6f03ecacd3870603766357505affd671ae
Nice work. I guess since your work integrates better with the existing derivations we should use that one. Would you mind bumping to 10.1.243 and making a PR?