feat: missing block catchup

[segfault-magnet]

closes: #23

The committer has been altered so that if we find that we haven't tried to commit the latest epoch-divisible block, we will fetch and schedule its commitment.

Known caveats:

• Will only commit blocks if there were no registered attempts to commit it before. (If the previous attempt failed because of network issues, it won't be retried)
• In the case of the db being deleted, upon starting it will attempt to commit the latest epoch-divisible block (wasting gas if it was already committed)

Fixing 1. would require separating errors into transient/nontransient and committing if previously the error was transient. Also setting some sort of retry limit per block.

Fixing 2. would require us to add a startup check in case of an empty db that would use the read-only query methods of the fuel state contract to determine whether the latest epoch-divisible block had already been committed.

If you feel either 1 or 2 to be of importance drop a comment and I'll make it into a task.

[Voxelot]

The tests look really nice and readable, nice work!

[Voxelot]

issue 2 should probably be solved for (as a followup), as the contract doesn't perform any checking for redundant commits which would reset the finality of the latest epoch block:

https://github.com/FuelLabs/fuel-bridge/blob/main/packages/portal-contracts/contracts/fuelchain/FuelChainState.sol#L89