treewide: prepare Rust packages for structuredAttrs by putting variables in env

[SFrijters]

These are necessary but possibly insufficient fixes: moving environment variables into env.{ ... }.

Tested with __structuredAttrs = true; for cargo, clippy, and rustfmt.

Updated the other packages found by grepping for RUSTC_BOOTSTRAP.

Should be a zero rebuild scenario, I hope?

Things done

• Built on platform:
  • x86_64-linux
  • aarch64-linux
  • x86_64-darwin
  • aarch64-darwin
• Tested, as applicable:
  • NixOS tests in nixos/tests.
  • Package tests at passthru.tests.
  • Tests in lib/tests or pkgs/test for functions and "core" functionality.
• Ran nixpkgs-review on this PR. See nixpkgs-review usage.
• Tested basic functionality of all binary files, usually in ./result/bin/.
• Nixpkgs Release Notes
  • Package update: when the change is major or breaking.
• NixOS Release Notes
  • Module addition: when adding a new NixOS module.
  • Module update: when the change is significant.
• Fits CONTRIBUTING.md, pkgs/README.md, maintainers/README.md and other READMEs.

---

Add a 👍 reaction to pull requests you find important.

[SFrijters]

nixpkgs-review result

Generated using nixpkgs-review.

Command: nixpkgs-review pr 472376
Commit: 6712843d3a748f71447dcb1c54ede3dfe9a42a76

[Mindavi]

Looks good and will help others prepare the same way when copying from one of these.

[wolfgangwalther]

Diff LGTM

[vcunat]

As discussed on chat, this is a mass rebuild (about 50k jobs on Hydra). No idea why it hasn't been detected, but except for x86_64-linux it rebuilds cargo via introducing empty RUSTFLAGS.

[vcunat]

I was re-checking it by commands like

nix eval -f. cargo.outPath --argstr system aarch64-linux

on the merge commit and on its first parent.

[wolfgangwalther]

Thanks for reverting, @SFrijters you'll want to target staging with this again.

It seems like CI didn't run to completion on this one (?), at least there is a red X on the last commit somehow. Not sure why it was possible to merge and why it was labeled with rebuilds and... @NixOS/nixpkgs-ci you might be interested in this.

[vcunat]

I've already notified https://matrix.to/#/#nixpkgs-ci:nixos.org

[vcunat]

I suspect the timing of force-pushes played a role.

[wolfgangwalther]

It seems like CI didn't run to completion on this one (?), at least there is a red X on the last commit somehow.

Actually that red X was from a re-run triggered by @vcunat after the PR was already merged. Looking at the original run, everything passed.

The labels were 3 weeks old. I suspect that this only turned into a mass-rebuild within the last few weeks? That's the most likely explanation.

[vcunat]

Looking at the UI here, the last push happened 6 minutes before the labels got applied.

[wolfgangwalther]

Looking at the UI here, the last push happened 6 minutes before the labels got applied.

Which is exactly the time that CI needs to run, so that's perfectly fine. The very first push didn't apply labels, because CI failed on that commit, so they only appeared after the second (and last) push. But as I said earlier, this was 3 weeks ago, so the mass rebuild was introduced on master inbetween.

[SFrijters]

Sorry for the mess - I do blindly follow CI for systems I can't build myself.

Just now I wanted to create a new branch/PR for staging and cherry-pick the commits on there, but they are all already on staging. Did master get merged into staging before this PR got reverted on master?

[SFrijters]

As discussed on chat, this is a mass rebuild (about 50k jobs on Hydra). No idea why it hasn't been detected, but except for x86_64-linux it rebuilds cargo via introducing empty RUSTFLAGS.

Not trying to nit-pick but just checking that we're on the same page: the effect is that an existing RUSTFLAGS = null; gets removed, right?

[vcunat]

Did master get merged [...]

Now I merged master into staging manually, so I think you should be good.

[vcunat]

Generally null values get ignored but empty strings get defined as empty existing env vars.