fetchurl: mitigate the change of pname/version special value

[ShamrockLee]

Mitigate the out-of-tree breakage caused by backward incompatible commit 1ec0227 ("fetchurl: don't treat empty-string name special") from PR #455994 by nullifying empty pname and version for Nixpkgs 25.11.

This PR depends on:

• fetchurl: make pname and version overridable via <pkg>.overrideAttrs #458183

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.

[ShamrockLee]

nixpkgs-review result

Generated using nixpkgs-review.

Command: nixpkgs-review pr 458150
Commit: 29153c27e4726166ef6a6ef6e2fd5a2147b2f32a

[philiptaron]

Can you cite an out-of-tree user that was broken?

[ShamrockLee]

Can you cite an out-of-tree user that was broken?

Issue oxalica/rust-overlay#239 reports the breakage, and PR oxalica/rust-overlay#241 fixes/works around it.

[philiptaron]

What this is saying is that:

• if lib.trivial.release is 25.05 or older, then add no attrs
• if lib.trivial.release is 25.11 or newer, then:
  • if pname is the empty string, set pname to null
  • if version is the empty string, set version to null

As I read it, this becomes a permanent addition, rather than something that disappears once we're past 2505.

[ShamrockLee]

Isn't !(lib.oldestSupportedReleaseIsAtLeast 2511) false on 26.05 and true on 25.11 and 25.05?

[philiptaron]

nix-repl> lib = import ./lib

nix-repl> lib2511 = lib.extend (self: super: { trivial = super.trivial // { oldestSupportedRelease = 2511; }; })

nix-repl> lib2605 = lib.extend (self: super: { trivial = super.trivial // { oldestSupportedRelease = 2605; }; })

nix-repl> !(lib.oldestSupportedReleaseIsAtLeast 2511)
true

nix-repl> !(lib2511.oldestSupportedReleaseIsAtLeast 2511)
false

nix-repl> !(lib2605.oldestSupportedReleaseIsAtLeast 2511)
false

[ShamrockLee]

IIUC, the current lib will go to the release-25.11 release branch, lib2511 will be what is on release-26.05, and lib2605 on release-26.11. Is that right?

[philiptaron]

My understanding:

• lib will be in master until ~Jan 2026, and will be in release-25.11 for its lifetime
• lib2511 is what will be on master starting ~Jan 2026 through til ~July 2026, and will be in release-26.05 for its lifetime.
• lib2605 is what will be on master starting ~Jul 2026, and will be in release-26.11 for its lifetime.

So yes, but also consider the master branch as well.

[ShamrockLee]

Can you cite an out-of-tree user that was broken?

Issue oxalica/rust-overlay#239 reports the breakage, and PR oxalica/rust-overlay#241 fixes/works around it.

In this case, an if-else or lib.optionalAttrs would be enough or even more elegant. Still, if the condition is from finalAttrs, dynamically decide whether to pass name would result in infinite recursion, and it would be important to know the default value.

Oh wait, fetchurl didn't support finalAttrs before #455994, so every such cases should be able to work around with conditional pass of name-related attributes if we don't backport commit 1ec0227 in PR #455994. On the other hand, having this compatibility layer will allow us to backport finalAttrs support with at least one consistent default name (null).