lib.types: remove __attrsFailEvaluation

[philiptaron]

Description of changes

The test (nix-build pkgs/test/release/default.nix) continues to pass without this preventative measure.

This line was added in #269356.

Things done

• Tested, as applicable:
  • nix-build pkgs/test/release/default.nix
• Fits CONTRIBUTING.md.

[lf-]

sorry, I'm done with the nix organization. can you request someone else for reviews? thanks!

[roberth]

Hi @lf-, is there something I could do?
Organization was always an overstatement, so I like to focus on achieving things together with others in the community.

@philiptaron
As for these PRs you made, they seem to be part of an incomplete effort by someone who's left GitHub, so we could either finish that work or remove it. Until someone wants to lead such an effort, I think it's better to keep the code clean, and it's easy enough to re-add something like this when needed.

[lf-]

nope, sorry. #323887 I've done my best to make things work but at some point you just have to call it for the health of all parties involved.

[philiptaron]

As for these PRs you made, they seem to be part of an incomplete effort by someone who's left GitHub, so we could either finish that work or remove it. Until someone wants to lead such an effort, I think it's better to keep the code clean, and it's easy enough to re-add something like this when needed.

Thanks for merging. The ghost was (as you likely know) @amjoseph-nixpkgs, and the work was attempting to make the eval checks faster. I think this is a good goal, and the test (pkgs/test/release/default.nix) means it could be checked in CI. I see no evidence that it currently is, but I definitely could be missing something.

The set of PRs is motivated by the same sentiment you expressed -- "it's better to keep the code clean" -- and if the test passes with the code removed, all the better.