writeShellApplication: use shellcheck only where supported

[fgaz]

Description of changes

GHC (=> shellcheck) isn't supported on some platforms (such as risc-v)
but we still want to use writeShellApplication on those platforms

Things done

• Built on platform(s)
  • x86_64-linux
  • aarch64-linux
  • x86_64-darwin
  • aarch64-darwin
• For non-Linux: Is sandbox = true set in nix.conf? (See Nix manual)
• Tested, as applicable:
  • NixOS test(s) (look inside nixos/tests)
  • and/or package tests
  • or, for functions and "core" functionality, tests in lib/tests or pkgs/test
  • made sure NixOS tests are linked to the relevant packages
• Tested compilation of all packages that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage
• Tested basic functionality of all binary files (usually in ./result/bin/)
• 23.11 Release Notes (or backporting 23.05 Release notes)
  • (Package updates) Added a release notes entry if the change is major or breaking
  • (Module updates) Added a release notes entry if the change is significant
  • (Module addition) Added a release notes entry if adding a new NixOS module
• Fits CONTRIBUTING.md.

[nixos-discourse]

This pull request has been mentioned on NixOS Discourse. There might be relevant details there:

https://discourse.nixos.org/t/prs-ready-for-review/3032/2401

[neilmayhew]

Wouldn't it make more sense to tryEval shellcheck itself?

[fgaz]

I tried, but it returns true, same for non-boot ghc. I think it's because the platform support in the ghc expression is implemented with throw

[Artturin]

something like this

Suggested change

	shellcheckSupported = (builtins.tryEval ghc.bootPkgs.ghc).success;
	shellcheckSupported = lib.meta.availableOn stdenv.hostPlatform ghc;

or

Suggested change

	shellcheckSupported = (builtins.tryEval ghc.bootPkgs.ghc).success;
	shellcheckSupported = lib.meta.availableOn stdenv.hostPlatform shellcheck;

[fgaz]

Don't you mean buildPlatform? Also this evaluates to true. I think it's because the platform support in the ghc expression is implemented with throw

[SuperSandro2000]

Can we convince the ghc people to change that to meta.platforms? Maybe it is super simple because they generate many things.

[Artturin]

@NixOS/haskell

[sternenseemann]

All the checks proposed here are (slightly) incorrect. That using meta.platforms is not possible is a slight oversight, namely that (shallowly) evaluating passthru (which mkDerivation needs to do) forces binDistUsed. I'll correct that separately and then it should be possible again to use meta.platforms.

The correct check for shellcheck's availability is lib.meta.availableOn stdenv.buildPlatform shellcheck.compiler which checks if the compiler we would use to build shellcheck itself is available. I need to warn you though that meta.platforms is imperfect, so there may be some edge cases, but I'm not sure if this applies to any platforms stdenv supports as a build platform. This is improved in #212188.

Edit: #246805

[fgaz]

Ping

[sternenseemann]

Note that the current approach will stop working after #243619. Seems like we will need to invent a completely new approach for this sort of thing then (i.e. some way to check if GHC can be bootstrapped in the context of the current package set(s)).

[sternenseemann]

#339272 will break this. Probably no other option than using tryEval then.