Missing optimization with transmute_copy on enums

[Jarcho]

Similar to #109958, but with transmute_copy. Transmuting a fieldless enum to it's integer type causes the optimizer to lose track of it's possible values.

Example:

#[repr(u32)]
enum E { X = 0 }
fn f(x: E) {
    assert_eq!(unsafe { *core::mem::transmute_copy::<E, i32>(&x) }, 0); // Not optimized out
}

cc Lokathor/bytemuck#175

[scottmcm]

One possible "fix" would be to stabilize some kind of wrapper around https://doc.rust-lang.org/nightly/std/intrinsics/fn.transmute_unchecked.html, because arguably code like this never actually wanted to use transmute_copy, and if it were to use transmute_unchecked instead then it would optimize: https://rust.godbolt.org/z/PWW4TE8s3