haskell.compiler.ghc{924,963,984}Binary: use broken for NCG check

[emilazy]

Fix availability checks on platforms where this would be false.

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.

[emilazy]

Wonder if it’d be good to have an advisory CI eval check for some platform that we don’t fully support but that you could reasonably expect to at least have working eval (e.g. one of i686, ARMv7, RISC‐V, LoongArch64).

[wegank]

LGTM.

Anecdote: In loongson-community@0d111ea, @darkyzhou added LoongArch support for GHC 9.8 by bootstrapping from Debian binaries (while LoongArch only has NCG backend support in 9.14+). I assume we'll remove the broken condition downstream during the next loong-testing branch rebase, but it's been a while since the last rebase, anyway.

[emilazy]

Looks like my work in #440774 should help a lot with that; funny that someone else was already backporting LLVM support patches to GHC.

Interesting to hear that the Debian binary package works. Maybe @sternenseemann talking about other distros’ binaries not working is out of date. Not sure whether that’s more or less nice than the cross‐built GHC 9.4.8 bootstrap plan, but it’s at least fewer moving parts. I do wonder how Debian produced those binaries in the first place.

I don’t think the broken condition will need adjusting once #440774 lands, since ghc966LoongArch64Binary will be a separate package. Actually, it should simplify things, since most of the GHC patches will already be incorporated by my PR.

It would be cool to see some of this stuff upstreamed, especially if we wire up the “chain everything back to GHC 9.4.8” plan.