issue in `inline assembly` part: `Late output operands`

[PeterWrighten]

In Late output operands part, the code snippet showed

If inlateout was used, then a and c could be allocated to the same register, in which case the first instruction to overwrite the value of c and cause the assembly code to produce the wrong result.

But when I implement the code

#![allow(unused)]
fn main() {
use std::arch::asm;

let mut a: u64 = 4;
let b: u64 = 4;
let c: u64 = 4;
unsafe {
    asm!(
        "add {0}, {1}",
        "add {0}, {2}",
        inlateout(reg) a, // change inout into inlateout here
        in(reg) b,
        in(reg) c,
    );
}
assert_eq!(a, 12);
}

, which changes inout into inlateout, the assembly code still produces correct result, and nothing wrong happens, that's weird and can't match what it declared.
So I think maybe asm! macro has been updated and stable, inlateout keyword has been entended or just that example is uncertain to describe such bad case.

[PeterWrighten]

I got what it means...I think it should declare explicitly that inlateout would produce wrong results in release mode...