fix(benches): Fixes compiler errors for benches

[mw2000]

We fix compiler errors that we see while benchmarking.

[mw2000]

A few runtime errors that are still to be fixed, which I wanted some feedback on -

Tests:

I see these tests failing - consensus::tests::test_verify_finality_invalid_finality and consensus::tests::test_verify_update_invalid_finality. Both fail at thread 'consensus::tests::test_verify_update_invalid_finality' panicked at ethereum/src/consensus.rs:793:55: and thread 'consensus::tests::test_verify_finality_invalid_finality' panicked at ethereum/src/consensus.rs:836:64: respectively.

They fail because the verify_finality_update and verify_update calls don't result in errors. How can we make these calls fail so we can pass the tests?

---

Benchmarks:
Benchmarking also seem to fail at get_balance

Benchmarking get_balance: Warming up for 3.0000 sthread 'main' panicked at benches/get_balance.rs:35:50:
called `Result::unwrap()` on an `Err` value: out of sync: 1730275786 seconds behind

Location:
    /Users/<user>/Desktop/helios/core/src/client/node.rs:60:9
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

[mw2000]

Believe @0xRampey is also working on the compiler errors within tests. Since he found the issue first, we should merge his PR in. I'll reshape this PR to target the benches crate while his PR targets tests

[0xRampey]

They fail because the verify_finality_update and verify_update calls don't result in errors. How can we make these calls fail so we can pass the tests?

Turns out thatLightClientHeader::default() when passed as finalized header, converts to None along the way and slips through verify_generic_update without being handled! Hence, the calls are passing.
I've added a new Error to handle that case and replaced ::default() header with those from older updates. That seems to be doing the trick.

[ncitron]

LGTM!