Ethereum test failure

[kayabaNerve]

I'm really unsure why this failed.

The tests passed in 30a77d8 and failed in 25c02c1.

The only Cargo.lock changes bumped ed25519-dalek, zeroize, and rustls in db49a63 (none should have impact here).

crypto and coins were not further edited. common only had serai-db edited.

---- test_ecrecover_hack stdout ----
thread 'test_ecrecover_hack' panicked at coins/ethereum/tests/contract.rs:123:48:
called `Result::unwrap()` on an `Err` value: Contract call reverted with data: 0x08c379a00000000000000000000000000000000000000000000000000000000000000020000000000000000000000000000000000000000000000000000000000000001065637265636f766572206661696c656400000000000000000000000000000000

Location:
    /home/runner/work/serai/serai/coins/ethereum/src/contract.rs:21:6

That appears to be far too many zeroes to be valid, if that is the calldata. That'd suggest a bug re: zeroization, yet ethereum-serai doesn't pull in zeroize (at least, not directly).

The tests pass locally. I'll rerun them on the CI now.

[kayabaNerve]

The CI consistently fails. It uses the nightly foundry which updated k256 in foundry-rs/foundry#6327. k256 changed their ecrecover functionality, as documented in RustCrypto/elliptic-curves#988, which borked revm. foundry uses revm. Not our fault.

[kayabaNerve]

foundry-rs/foundry#6358 will close this.

[kayabaNerve]

We got affected by this due to #116. It seems while foundry does not have stable versions to pin to, you can pin to nightly versions (I didn't realize you could pin to "nightly-x" instead of just "nightly") per foundry-rs/foundry-toolchain#35. Accordingly, we can resolve this (and #116) on our end.