Compiler generates suboptimal code when array is iterated (works fine with slice).

[VorfeedCanal]

Couldn't find if that was reported yet.

This is example from reddit:

This produces suboptimal results:

    for name in name_arr {
        print!("{}", name.get_name());
    }

while this works as expected:

    for name in &name_arr {
        print!("{}", name.get_name());
    }

Godbolt reference with comparison: https://godbolt.org/z/1T6G3s7Gz

[nikic]

Should be fixed in LLVM 16, assigning to track.

[nikic]

Fixed by the LLVM 16 upgrade.

[nikic]

Not going to add a separate test for this, I think this is effectively covered by #101082.