gcc backend: remove duplicate handling for max/minimumf128

[RalfJung]

minimumf128 and maximumf128 are for some reason handled both in get_simple_intrinsic and get_simple_function. Either way they are actually called like functions though, not GCC intrinsics, so let's keep the one one in get_simple_function.

r? @antoyo

[rustbot]

Some changes occurred in compiler/rustc_codegen_gcc

cc @antoyo, @GuillaumeGomez

[antoyo]

Please run the core tests locally to check this change since the CI doesn't run the core tests in here.

[RalfJung]

I don't have a setup to run GCC tests. I was hoping CI would take care of that...

[antoyo]

Unfortunately, the CI here do not run the core tests (it compiles core with LLVM).
Only the CI of the cg_gcc repo runs the core tests.

[RalfJung]

Ah, that's good; I will re-send the PR there.

[RalfJung]

This code already doesn't exist any more in that repo, it got removed in rust-lang/rustc_codegen_gcc@5b2cfe3 -- seems like there just hasn't been a sync since then.

[antoyo]

This code already doesn't exist any more in that repo, it got removed in rust-lang/rustc_codegen_gcc@5b2cfe3 -- seems like there just hasn't been a sync since then.

Oh.
The sync is currently blocked on a ThinLTO issue.
I'm currently looking at this with Bjorn: hopefully, that should be fixed in the following days.

[RalfJung]

That's a sync in the other direction, but I guess that has to happen first or so.