feat: configurable e2e setup

[segfault-magnet]

related to: #84
closes: #86

To run e2e tests from now on you have to have fuel-core and foundry in your PATH.

The e2e test rn:

    async fn submitted_correct_block_and_was_finalized() -> Result<()> {
        // given
        let show_logs = true;
        let stack = WholeStack::deploy_default(show_logs).await?;

        // when
        stack
            .fuel_node
            .client()
            .produce_blocks(stack.deployed_contract.blocks_per_commit_interval())
            .await?;

        // then
        stack
            .committer
            .wait_for_committed_block(stack.deployed_contract.blocks_per_commit_interval() as u64)
            .await?;
        let committed_at = tokio::time::Instant::now();

        sleep_until(committed_at + stack.deployed_contract.duration_to_finalize()).await;

        let latest_block = stack.fuel_node.client().latest_block().await?;

        let validated_block =
            BlockValidator::new(stack.fuel_node.consensus_pub_key()).validate(&latest_block)?;

        assert!(stack.deployed_contract.finalized(validated_block).await?);

        Ok(())
    }

How it works

Build phase

In the build phase of e2e we're downloading a zip of fuel-bridge and extracting the FuelChainState.sol.

We then initialize a foundry project, add the required deps and compile the contract.

The artifacts are stored in the target directory. If you change the revision of fuel-bridge used or do a cargo clean this process is repeated.

Usage in tests

Each service is now a structure. Running an eth node (currently anvil from foundry) and deploying the contract looks like this:

        let eth_node = EthNode::default().with_show_logs(logs).start().await?;
        let deployed_contract = eth_node
            .deploy_contract(
                finalize_duration.as_secs(),
                blocks_per_interval,
                commit_cooldown_seconds,
            )
            .await?;

If you drop eth_node the node will shut down.

The remaining services are done similarly.

Contract deployment had to be done via a solidity script using forge script because the smart contract needs to be accessed via a proxy contract (because of how the contract is written you cannot get the role needed for committing if you deploy only the contract).

Wallet keys, contract addresses and the fuel-core consensus keys are generated on each run and no longer hard coded.

[Br1ght0ne]

Solid work!

[segfault-magnet]

Task here

[segfault-magnet]

Task here

[digorithm]

Beautiful work. Can't thank you enough for dropping the need for docker and docker compose to run these tests. Massive win, IMO.

[hal3e]

Awesome work!