fixed-hash: Derive PartialEq and Eq instead of implementing them manually

[fhartwig]

The motivation for this is that Rust only lets you use named constants (as opposed to literals) for pattern matching if the value's type has derived PartialEq and Eq implementations. So trying to compile the following code:

use ethereum_types::H160;

const ONE: H160 = H160::repeat_byte(1);;
const TWO: H160 = H160::repeat_byte(2);

fn main() {
    let x = H160::repeat_byte(3);
    match x {
        ONE => println!("it's a one!"),
        TWO => println!("it's a two!"),
        _ => println!("something else"),
    }
}

currently results in the following compilation error:

error: to use a constant of type `H160` in a pattern, `H160` must be annotated with `#[derive(PartialEq, Eq)]`
 --> src/main.rs:9:9
  |
9 |         ONE => println!("it's a one!"),
  |         ^^^

error: to use a constant of type `H160` in a pattern, `H160` must be annotated with `#[derive(PartialEq, Eq)]`
  --> src/main.rs:10:9
   |
10 |         TWO => println!("it's a two!"),
   |         ^^^

As far as I can tell, the existing explicit PartialEq implementation exists for historical reasons and isn't necessary any more, the derived implementation should be equivalent. Assuming that is true, the only potential drawback of this change that I can see is that it poses a backwards compatibility hazard, i.e. if in the future you would want to manually implement PartialEq again for some reason, doing so would break the code of anyone using e.g. named H160 constants in pattern matches.

[ordian]

Thanks for the PR!

Could you add a test similar to the one introduced in #421?

[fhartwig]

Sure, done!

[ordian]

Thanks! Only cargo +nightly fmt is missing: https://github.com/paritytech/parity-common/actions/runs/4769435148/jobs/8484483754?pr=742

[fhartwig]

Sorry about that, fixed.