Our great sponsors
-
WorkOS
The modern identity platform for B2B SaaS. The APIs are flexible and easy-to-use, supporting authentication, user identity, and complex enterprise features like SSO and SCIM provisioning.
-
-
-
InfluxDB
Power Real-Time Data Analytics at Scale. Get real-time insights from all types of time series data with InfluxDB. Ingest, query, and analyze billions of data points in real-time with unbounded cardinality.
I myself use Fedora Silverblue with a https://ublue.it/ -based custom image, and I use home-manager for shell configuration. Sure, my GNOME layout isn't declarative, but basically everything else is. Pair that with one of the best NVIDIA driver experiences and the strong feeling of stability, it's better than most other Linuxes rn.
> When you have to patch things at the level of ELF, your basic architecture needs work.
May I ask why? It might sound uncommon, but that doesn't mean it's wrong. Nix only patches pre-built binaries targeting other distros. It patches the expected path of shared libraries so that the patched binaries can use versions of libraries that work best.
This is not even a Nix thing. It's standard practice when packaging pre-built binaries. You can see this for yourself in Debian code search:
https://codesearch.debian.net/search?q=patchelf&literal=1
All Nix does is automate the process that other distros end up doing anyways.
> The project management of Nix feels like a disaster
That's not a fair take when the reason for that is a single issue you care about not being fixed fast enough. Nix has merged 41 PRs in the past month alone, so it's not like there hasn't been any progress. The devs are doing the best they can with the resources available.
https://github.com/NixOS/nix/pulse/monthly
> May I ask why?
Because it breaks the universe?
https://github.com/NixOS/nixpkgs/issues/24744
https://discourse.nixos.org/t/using-mold-as-linker-prevents-...
And presumably this rattles into glibc vs musl, too.
patchelf is a working hack short term, but what's the long term solution? There is no long term plan as far as I can tell.
> That's not a fair take when the reason for that is a single issue you care about not being fixed fast enough. Nix has merged 41 PRs in the past month alone, so it's not like there hasn't been any progress. The devs are doing the best they can with the resources available.
That's just the most visible to newbies and one of the most popular of the errors. When a Nix tutorial has to go out of it's way to point out the inscrutable error message as well (see: https://fasterthanli.me/series/building-a-rust-service-with-...) and it's still not fixed--well, that speaks volumes.
When flakes can't get either killed, fixed, or stabilized after how long, that speaks volumes.
Nix has more than a few deficiencies. Nix flakes also has a quite a few deficiencies. And the most recent announcement just simply kicks the can down the road.
The problem is that project management doesn't seem to be able to corral things sufficiently to make progress rather than just gaining more problems on a rewrite (flakes and the new command line).
I mean, if a simple error message change can sit open for more than a year, is there really any hope for fixing anything truly substantial?