Fix test build.

[afck]

No description provided.

[afck]

#190 applies to aura-pos as well, not just to hbbft.
This PR only fixes the build, but the tests still fail.

[afck]

The authority_round test in ethcore/sync/src/tests/consensus.rs broke in the malice reports commit: ae01f9b

---- tests::consensus::authority_round stdout ----
thread 'tests::consensus::authority_round' panicked at 'assertion failed: `(left == right)`
  left: `1`,
 right: `2`', ethcore/sync/src/tests/consensus.rs:74:2
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace.

[vkomenda]

Does also parity-ipfs-api fail to compile? I have

error[E0433]: failed to resolve: could not find `format_args` in `core`
   --> ipfs/src/lib.rs:133:8
    |
133 |         _ => format!("{}:{}", interface, port),
    |              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ could not find `format_args` in `core`
    |
    = note: this error originates in a macro outside of the current crate (in Nightly builds, run with -Z external-macro-backtrace for more info)

[vkomenda]

Maybe the failure in parity-ipfs-api is a nightly glitch? I'll try a different nightly...

[afck]

No, I actually fixed it. Will push in a few minutes. (If I don't hit even more errors in the meantime…)
The problem here was that something was imported as core, which breaks the format! macro.

I also fixed tests::consensus::authority_round, which was my fault: Seeing that AuthorityRound::step is only used in tests, I marked it #[cfg(test)]. That was a mistake because:

1. Engine::step has a no-op default implementation.
2. Dependent crates use step in their own tests, which means cfg(test) is false! For that purpose there's the test-helpers feature. So I guess I should have marked fn step in all three places (the Engine trait, and the impls for AuthorityRound and Clique) as #[cfg(any(test, feature = "test-helpers))] or something.

But I'll just revert my change, and possibly suggest that upstream later.

[afck]

I added some fixes, but there's still a lot to do:

• v1::tests::eth::eth_get_block_by_hash still fails.
• I couldn't fix evmbin/benches, so I disabled it for now. It looks horribly outdated.

[vkomenda]

Another problem is that a dependent crate fails to compile as well - jsonrpc. It doesn't make sense to correct old versions of other crates. Maybe those problems are fixed in the latest stable, in which case we could rebase on it?

[vkomenda]

All ethcore tests pass, which is the main thing. There are problems outside ethcore due to the code becoming outdated. Maybe we could rebase onto the latest stable instead of starting to fix failures one by one.

[afck]

I agree! I'll take a look at the upstream branches tomorrow, and see by how many of these bugs they are still affected, then I'll either merge this PR, or we'll rebase.

[vkomenda]

The build errors in parity-ethereum only were fixed in openethereum#10991.

[vkomenda]

All tests pass if compiled with stable Rust 1.37.