lib/minver: bump to 2.3.17

[mweinelt]

The first version that supports zstd compression, to create the option to eventually switch compression for the binary cache.

It was released one year ago on 2023-11-03 and first shipped in NixOS 23.11.

Things done

• Built on platform(s)
  • x86_64-linux
  • aarch64-linux
  • x86_64-darwin
  • aarch64-darwin
• For non-Linux: Is sandboxing enabled in nix.conf? (See Nix manual)
  • sandbox = relaxed
  • sandbox = true
• Tested, as applicable:
  • NixOS test(s) (look inside nixos/tests)
  • and/or package tests
  • or, for functions and "core" functionality, tests in lib/tests or pkgs/test
  • made sure NixOS tests are linked to the relevant packages
• Tested compilation of all packages that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage
• Tested basic functionality of all binary files (usually in ./result/bin/)
• 24.11 Release Notes (or backporting 23.11 and 24.05 Release notes)
  • (Package updates) Added a release notes entry if the change is major or breaking
  • (Module updates) Added a release notes entry if the change is significant
  • (Module addition) Added a release notes entry if adding a new NixOS module
• Fits CONTRIBUTING.md.

---

Add a 👍 reaction to pull requests you find important.

[winterqt]

Technically this is a breaking change, but I think this is fine -- realistically, nobody is still evaluating Nixpkgs with 2.3.16 or below.

I can see arguments for and against a changelog entry, given we're just bumping to a patch version, so that's up to you.

[mweinelt]

Changelog added.

[winterqt]

Thanks! LGTM. cc @RossComputerGuy for final OK as 24.11 RM.

[RossComputerGuy]

[winterqt]

Apparently I can't read and didn't notice @infinisil was pinged lol. Will wait on him.

[infinisil]

Hmm, this is not really what minver is supposed to be used for, it's documented as the "minimum required version for evaluating Nixpkgs". It's arguably even possible to evaluate with Nix 2.3 but substitute with 2.3.17 to use zstd. So is this really the right place to make this change? What happens if we just don't do this? Is there another mechanism that fits better?

Nothing obvious comes to mind, so this might really be the best way of doing it, but I'd like to see a bit more background on why this is needed.

[mweinelt]

That's fair.

Earlier versions will be unable to substitute, which makes nixpkgs a fairly expensive proposition. The solution is then to upgrade nix, and this kinda achieves the same effect, even though it does not break eval.

That being said, I'm fine with not doing this change, if that means we don't have to wait for minver changes to switch the compression algorithm.

[infinisil]

How about we make sure that all versions of Nix that don't support zstd are marked as deprecated in at least supported releases.

[mweinelt]

What mechanism do you have in mind?

[infinisil]

I had lib.warn in mind, but that messes up nix-env :/

[infinisil]

Hacky alternative that probably breaks other things: Wrap /bin/nix to print a deprecation warning to stderr before executing the actual Nix binary.

[infinisil]

Maybe this bump is the best way after all, because the above options just suck. We'd need to fix nix-env really, but that's a huge effort.

[winterqt]

Yeah, I'd be fine with merging if it's the only good option. (curses nix-env from the distance)

[Mindavi]

I think this is the best way to express that. People using older versions can indeed eval, but shouldn't expect the cache to work. Which is one of the things people find very valuable, I think.