devenv: 2.3.0 -> 2.3.1

[domenkozar]

Patch release of devenv. Changelog: https://github.com/cachix/devenv/releases/tag/v2.3.1

Only the version, source hash and cargoHash change. The pinned cachix/nix revision is unchanged.

Disclosure: this pull request summary and the commit were produced with Claude Code (Claude Fable 5.1), reviewed by me.

Things done

• Built on platform:
  • x86_64-linux
  • aarch64-linux
  • 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.
• Follows the automation/AI policy.

[RossSmyth]

Please follow the PR guidelines, do not delete them, or as I suspect in this instance it was completely ignored and the text seen here is just whatever the llm calculated based upon its training windows. This PR blatantly does not follow contribution guidelines. Please reopen when it does.

[domenkozar]

Never seen you interested into devenv, but it's good you care :)

Could you explain which rule I've violated?

[fzakaria]

This a pretty mundane change.

I'm also genuinely curious why this was closed or what was the missing/incorrect part?

[NyCodeGHG]

Please read the Automation Policy.
You are not using the PR template, it's missing the "I read the Automation Policy Checkbox" and it still lists x86_64-darwin.
You also did not correctly mark your commit as AI-assisted according to the Automation Policy.

[jackrosenberg]

Why require Assisted-by: over Co-authored-by:?

Since LLM‐based AI tools use Co-authored-by: attribution by default, we think that using a different format will serve as a brown M&M test and assist with triage. A commit that contains such a tool in Co-authored-by: but not Assisted-by: is, by definition, one that did not follow the policy and can be closed, potentially automatically. It also matches the format standardized on by the Linux kernel and Fedora, and will make it easier to review and query for use of these tools.

• #514587

The disregard for following guidelines shows a concerning level of apathy.

[SigmaSquadron]

(As mentioned above. This change request is just to prevent an accidental merge.)

[vcunat]

Honestly, the change is trivial, so I believe it doesn't matter whether the hashes were copy&pasted manually or generated in any other way. Why add complicated hurdles in cases where the benefit doesn't really exist?

[SigmaSquadron]

Because its our agreed upon policy. We're not asking the author to cease using AI, despite its many issues in environmental, ethical, quality and legal domains. We're simply asking a minimal amount of respect for our existing guidelines.

[vcunat]

This was marked as generated with help of AI from the very start. (in the commit message)

[domenkozar]

What happened here is that I have a release process as part of devenv in that repository, and it didn't pick up CLAUDE.md from nixpkgs when doing the PR.

I'll fix that so that it respects the template and the commit format.

These changes are really trivial, of course you can pick up on me if you want, but it shows a lack of empathy for only trying to maintain my software.

Worth mentioning is that none of this behavior will make me stop using AI.

[SigmaSquadron]

This was marked as generated with help of AI from the very start. (in the commit message)

It was not marked in a way that was compliant with our pre-established guidelines.

Worth mentioning is that none of this behavior will make me stop using AI.

As I previously mentioned, no one here is trying to stop you from using whatever automation tools you deem appropriate. We're just asking you to follow the rules agreed upon by the community over the usage of said tools. The fact that you enjoy using these tools does not put you in a position to disregard our guidelines.

[SigmaSquadron]

Please note that the reviews still stand, as the co-authored-by tag is invalid per our policy.

[domenkozar]

Please note that the reviews still stand, as the co-authored-by tag is invalid per our policy.

Having a Co-authored-by trailer alongside Assisted-by is not prohibited anywhere in the text.

Resolved

[jackrosenberg]

Literally all they had to do was follow the guidelines.

[EpicEric]

@drupol To be clear, nobody argued against the LLM usage, the disagreement was that the pull request template and AI contribution guidelines were not followed.

[drupol]

✅ nixpkgs-review result

Generated using nixpkgs-review-gha (07e0eb6)

Command: nixpkgs-review pr 562437
Commit: f3f59c37723f46de21ace5e88dfcdfa63b6dabab (subsequent changes)
Merge: da5ca5203635381497d1dede9c50fc78204e73f0

Logs: https://github.com/drupol/nixpkgs-review-gha/actions/runs/34714661557/attempts/1

---

x86_64-linux

✅ 1 package built:

• devenv

---

aarch64-linux

✅ 1 package built:

• devenv

---

x86_64-darwin

✅ No rebuilds

---

aarch64-darwin (sandbox = relaxed)

✅ 15 packages built:

• cachix
• devenv
• haskellPackages.cachix
• haskellPackages.hercules-ci-agent
• haskellPackages.hercules-ci-cli
• haskellPackages.hercules-ci-cnix-expr
• haskellPackages.hercules-ci-cnix-store
• hci
• hercules-ci-agent
• nix-heuristic-gc
• nix-inspect
• nix-plugins
• nixVersions.nix_2_31
• omnix
• tests.haskell.cabalSdist.hercules-ci-cnix-store

[drupol]

Approved automatically following the successful run of nixpkgs-review.

[xokdvium]

Are we really going to apply patches to nix without getting a nix team sign off? It's a backport but still, wtf?

[domenkozar]

Feel free to revert it if you think there's something wrong with that patch.

It's there to unblock nixpkgs-unstable.

[xokdvium]

I'm not looking to revert, just please be kind and give a bit of grace for people to look at the stuff. If you'd prefer, that can also be backported in the nix tree so that we are not carrying patches around.

[domenkozar]

I'll make backport PRs in the nix repo tomorrow.

Sorry for the mess, I hope we unblocked nixpkgs and fixed some devenv regressions.

[vcunat]

the disagreement was that the pull request template and AI contribution guidelines were not followed

• AI usage had been disclosed from the very start in the commit message; yes, by a slightly different line than required by the contribution guidelines
• the template was fixed before the PR got closed (that's what I see in the GitHub's edit log above)

[vcunat]

Also, interestingly, Domen got basically abused on https://matrix.to/#/#dev:nixos.org in parallel with this PR's discussion. Personal attacks pass OK but these nits are against guidelines?

As for the guidelines, they require to "transparently disclose any non‐trivial use of automation". The discussion happened on a simple bump of version+hashes (4 lines), which I'd argue is quite trivial in this respect.

[domenkozar]

I've worked on this project for more than 10 years and I have no idea why I deserve this treatment.

Especially on PRs like this where we fix bugs that would need to be rolled out fast to address issues people face with OSS we give for free.

But okay, keep going with the hate, I hope it works out for you all.