Experiment with a single public input for the light client circuit.

[philippecamacho]

CAUTION: this PR is not meant to be merged. Its purpose is to evaluate how much gas is required when calling the state update function of the light client contract when the public input is of size 1, instead of 8.

Run the benchmark:

> cargo build --bin diff-test --release
> forge test --mt testCorrectUpdate

Running 1 test for contracts/test/LightClient.t.sol:LightClient_newFinalizedState_Test
[PASS] testCorrectUpdate() (gas: 851965)
Test result: ok. 1 passed; 0 failed; 0 skipped; finished in 16.06s
 

So misfortunately the gas cost does not decrease that much: 851,965 instead of 939,380.