cudaPackages.cuda-library-samples: Only enable on Linux, fix channel#279684
cudaPackages.cuda-library-samples: Only enable on Linux, fix channel#279684infinisil merged 1 commit intoNixOS:masterfrom
Conversation
This very weirdly broke the channel evaluation: https://hydra.nixos.org/build/245871962/nixlog/1 It appears that this attribute is only evaluated by Hydra, _not_ by ofborg. So this wouldn't have been detected by CI anyways in the PR that introduced the problem: NixOS#276800. However, due to NixOS#271123 (comment), the channel only broke once that was fixed with NixOS#278777 Whether the fix is good, I don't know, but the failing-on-darwin attribute doesn't exist anymore with this commit, making the tarball build succeed again: nix-build pkgs/top-level/release.nix -A tarball
lib.callPackageWith: Use abort again instead of throw and fix evaluation errors caused by it
#278777
There was a problem hiding this comment.
The more cursed that the linked PR didn't even touch cuda-library-samples directly (must've relaxes some filters at the scope level?)
The "package" is in a pretty horrible shape, almost unmaintained, doesn't set neither platforms nor badPlatforms. It's also somehwat un-packageable: the upstream uses recursive-gnumake which eventually calls cmake.
I also think that optionalAttrs is a good way to create more eval erros in the future, but if this fixes Hydra right now we should just merge
Thanks for the ping and for the fix
Yeah this was a pain to figure out. Apparently the files nixpkgs/pkgs/top-level/cuda-packages.nix Lines 116 to 117 in 2c9008a This is not great, it violates implicit Nixpkgs boundary: |
|
And only adding packages to a package set if they are supported on the platform is also very much not conventional. |
Tracking in #266475 (even if the title doesn't suggest)
Opened #279703. Not even sure how long they've been in |
This very weirdly broke the channel evaluation: https://hydra.nixos.org/build/245871962
It appears that this attribute is only evaluated by Hydra, not by ofborg. So this wouldn't have been detected by CI anyways in the PR that introduced the problem: #276800.
However, due to #271123 (comment), the channel only broke once that was fixed with #278777
Whether the fix is good, I don't know, but the failing-on-darwin attribute doesn't exist anymore with this commit, making the tarball build succeed again:
Ping @vcunat @NixOS/cuda-maintainers @amjoseph-nixpkgs @Mic92
Add a 👍 reaction to pull requests you find important.