buildGoModule: intersect given and available meta.platforms

[acid-bong]

At least Aerc and Syncthing aren't getting rebuilt, I can't remember any other Go-based programs on top of my head. Hopefully it's not gonna cause a mass rebuild

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]

I like the meta change!

The passthru change is reasonable, though I'm not 100% sure if it we need to add a release note for this change as it does touch the interface.

Here's a brief history of the current passthru design:

The design decision (I made in #225051 ) about the current passthru precedence predates the final version of lib.extendMkDerivation. At that time, I was still hoping that we could eventually pass everything directly into stdenv.mkDerivation (i.e., no removeAttrs or excludeDrvArgNames) and turn language- and framework-specific build helpers into attribute overlays one could load via overrideAttrs during package definition. (See comment #234651 (comment) for the motivation.)

I changed my mind after participating in the __structuredAttrs = true conversion. Backward compatibility can only be made possible with excludeDrvArgNames. Even stdenv.mkDerivation itself excludes env when __structuredAttrs is false. Comment #234651 (comment) explains this in detail.

[acid-bong]

Hmm, seems that non-overridable passthrus are more complicated. I'll remove that commit and leave only the meta part

[katexochen]

Could you explain again why we need this change? As I understand it, when I add a platform to meta.platforms that isn't supported by the go toolchain, this change will silently remove the unsupported platform from meta.platforms. Should we maybe throw a warning in this case, too?

[acid-bong]

Should we maybe throw a warning in this case, too?

sounds good, but it'll require changes in all builders for the sake of consistency