zstd: set meta.pkgConfigModules#436445
Conversation
I've chosen to use zstd from the pkgs fix point for the pkg-config test instead of using finalAttrs.finalPackage because this is effectively also done for all the other tests (they test bindings from the pkgs fix point which get their zstd from the pkgs fix point).
|
Indeed, didn't actually try. Seems like these just show up one by one how wonderful. |
|
@wolfgangwalther just needs #436468 |
| python-zstd = python3Packages.zstd; | ||
| haskell-zstd = haskellPackages.zstd; | ||
| haskell-hs-zstd = haskellPackages.hs-zstd; | ||
| pkg-config = testers.hasPkgConfigModules { package = zstd; }; |
There was a problem hiding this comment.
Same argument as in https://github.com/NixOS/nixpkgs/pull/436468/files#r2296679493, I think we should at least do the right thing for this test.
There was a problem hiding this comment.
These tests serve different roles. The reverse dependencies test the package set, whereas this should be a test for an intrinsic property of the package. If we can include overrides in the test, we should make it coherent like that. I'll merge this to unblock you, but this isn't quite right yet. |
I've chosen to use zstd from the pkgs fix point for the pkg-config test instead of using finalAttrs.finalPackage because this is effectively also done for all the other tests (they test bindings from the pkgs fix point which get their zstd from the pkgs fix point).
This will fix this regression on haskell-updates (since buildInputs is built based on
meta.pkgConfigModules(long story)).Things done
passthru.tests.nixpkgs-reviewon this PR. See nixpkgs-review usage../result/bin/.Add a 👍 reaction to pull requests you find important.