Removing match on constant

[tomusdrw]

Pattern matching on complex type constants without #[derive(PartialEq)] is being phased-out.

Details:
rust-lang/rfcs#1445

We cannot derive PartialEq because arrays does not implement any traits.

[tomusdrw]

Does this actually make sense? Before this refactor we were matching None twice:

None if self.code_cache.is_empty() => Some(&self.code_cache),
None => Some(&self.code_cache),

[arkpar]

Looks like it does not. None matches to Some(&self.code_cache) in all cases
@gavofyork please take a look, might be a bug in the original code.

[gavofyork]

the original code (https://github.com/ethcore/parity/blob/master/ethcore/src/account.rs) was:

Some(c) if c == SHA3_EMPTY  && self.code_cache.is_empty() => Some(&self.code_cache),
Some(_) if !self.code_cache.is_empty() => Some(&self.code_cache),
None => Some(&self.code_cache),
_ => None,

there are no options here which are superfluous.

[tomusdrw]

The code from master already contains this change (was included with #889 - otherwise there was compilation error on nightly).
This is what it looked like before:
https://github.com/ethcore/parity/blob/61420d3c9c64d87ab2c83d972c7d2d812b1d9b72/ethcore/src/account.rs#L141

The logic wasn't different, but there was superfluous | None in first branch.
My question is rather if when this method should return None - currently it's only the case when code_hash == SHA3_EMPTY && !code_cache.is_empty() OR code_hash != SHA3_EMPTY && code_cache.is_empty() - this is correct?

[gavofyork]

Correct (technically, the first is actually a logic error and should properly be either asserted against or an error). The second (i.e. there is definitely some non-empty code, but it isn't presently cached) properly returns None.

[gavofyork]

extra space.

[gavofyork]

needs to be resolved.