Integer hex literals should perhaps not consider signededness

[alexcrichton]

I would not expect this code to produce a warning:

fn main() {
    let a: i8 = 0x20;
    assert!(a == 0xff);
}

Although the following warnings are printed

foo.rs:3:12: 3:21 warning: comparison is useless due to type limits, #[warn(type_limits)] on by default
foo.rs:3     assert!(a == 0xff);
                     ^~~~~~~~~
<std-macros>:36:4: 52:5 note: in expansion of assert!
foo.rs:3:4: 3:23 note: expansion site
foo.rs:3:17: 3:21 warning: literal out of range for its type, #[warn(type_overflow)] on by default
foo.rs:3     assert!(a == 0xff);
                          ^~~~
<std-macros>:36:4: 52:5 note: in expansion of assert!
foo.rs:3:4: 3:23 note: expansion site

[lilyball]

Why shouldn't that produce a warning? Are hex literals somehow magic? I would expect to have to say either assert!(a == 0xffu8 as i8) or assert!(a as u8 == 0xff).

[alexcrichton]

Closing, general opinion is opposite of mine, I think I'm just special in this regard.