| #[repr(u8)] | |
| enum Eu8 { | |
| Au8 = 23, | |
| Bu8 = 223, | |
| Cu8 = -23, | |
| //~^ ERROR cannot apply unary operator `-` to type `u8` | |
| } | |
| #[repr(u16)] | |
| enum Eu16 { | |
| Au16 = 23, | |
| Bu16 = 55555, | |
| Cu16 = -22333, | |
| //~^ ERROR cannot apply unary operator `-` to type `u16` | |
| } | |
| #[repr(u32)] | |
| enum Eu32 { | |
| Au32 = 23, | |
| Bu32 = 3_000_000_000, | |
| Cu32 = -2_000_000_000, | |
| //~^ ERROR cannot apply unary operator `-` to type `u32` | |
| } | |
| #[repr(u64)] | |
| enum Eu64 { | |
| Au32 = 23, | |
| Bu32 = 3_000_000_000, | |
| Cu32 = -2_000_000_000, | |
| //~^ ERROR cannot apply unary operator `-` to type `u64` | |
| } | |
| // u64 currently allows negative numbers, and i64 allows numbers greater than `1<<63`. This is a | |
| // little counterintuitive, but since the discriminant can store all the bits, and extracting it | |
| // with a cast requires specifying the signedness, there is no loss of information in those cases. | |
| // This also applies to isize and usize on 64-bit targets. | |
| pub fn main() { } |