simulators/eth2/dencun: Dencun Interop Simulator and Builder Tests

[marioevz]

Changes included

• simulators/eth2/common
  • Refactors to better configure the different forks on the beacon chain, https://github.com/ethereum/hive/pull/838/files#diff-5bc8313f0b88dc95de0c3f4946671bd5333d8400f7e3d1664d3039956e407019, and the execution chain, https://github.com/ethereum/hive/pull/838/files#diff-0eb3479aaa196eba96e6861dc914cd2a4b1a3eb26e4ec7664dff30df2b7cb3eb
  • Updates to configure Deneb genesis state and config values: https://github.com/ethereum/hive/pull/838/files#diff-256d8428aa00506d07e2bc1aa1238d4ddd9d29479b9d34265fcdfeb30af243f9
  • Refactors to the genesis state generator constructs to more easily implement future forks (using a generic interface and implementation for each fork https://github.com/ethereum/hive/pull/838/files#diff-6cfaf4e1501b00a33cb7ef36176ab8fd327c586ea26795be386b86c2a98b0ff1)
  • Removed Proof-Of-Work related code: https://github.com/ethereum/hive/pull/838/files#diff-d2d18e03f0f8261bddbc2161aa12c9c3c78205ba79a83719863a172cd5d89cc2
  • Add Dencun Engine API versions of the endpoints: https://github.com/ethereum/hive/pull/838/files#diff-c46331fa03eada542f922f93b7a88fde365c836400d4047f8c5534a298f3e60b
• simulators/eth2/dencun
  • Dencun simulator now uses a suite interface that allows writing test suites more straightforward, and ideally we should inherit this construct for future forks: https://github.com/ethereum/hive/pull/838/files#diff-340fc224ae530ec3238ead0cc9a63f72e918f26fe0e7d158aec506b7fc7592ed
  • Four test suites have been created in the dencun sim:
    • Base: Testnet genesis and finalization, starting from either: Capella and fork to Deneb on Epoch 1; Deneb Genesis
    • Sync: Testnet genesis and sync of a client, starting from either: Capella and fork to Deneb on Epoch 1; Deneb Genesis
    • ReOrg: Testnet genesis to two different forked chains (with different blobs) and re-org to a single canonical chain
    • Builder: Testnet genesis, finalization, and builder workflow, starting from either: Capella and fork to Deneb on Epoch 1; Deneb Genesis
• simulators/eth2/*
  • Updates zrnt, eth2api eth-clients, mock-builder to the Deneb versions
    • zrnt still points to https://github.com/marioevz/zrnt/tree/update-deneb because the PR is unmerged
  • Fixes to engine, testnet and withdrawals simulators to account for the library updates

Linked Issues

• Closes Remove Pre-Merge Tests from simulators/eth2/engine, simulators/ethereum/engine #783

ToDos:

• Add proper verifications to blob transactions on the Execution chain and then on the beacon Beacon using the Beacon API
  • Verify KZG commitments on beacon block against blob sidecars
  • Cross-compare blob sidecars between multiple clients
  • Compare KZG commitments against execution payload header
• Add sync tests starting on Deneb with a chain that includes blobs
  • Start chain from Capella, upgrade to deneb, send blobs, wait an epoch, then start client to sync
  • Start chain from Deneb, send blobs, wait an epoch, then start client to sync
• Add Re-Org tests starting on Deneb with N nodes on M different chains containing different blobs, wait for clients to reach canonical chain consensus
  • Start chain from Capella, upgrade to deneb, send blobs, wait an epoch, then start client to sync
  • Start chain from Deneb, send blobs, wait an epoch, then start client to sync
• Add builder tests corrupting: blob gas used, excess blob gas

Examples

Run builder tests

cl=lighthouse; ./hive --client-file ./configs/cancun.yaml --client go-ethereum,${cl}-bn,${cl}-vc --sim eth2/dencun" --sim.limit "eth2-deneb-builder/"

Run sync tests

cl=lighthouse; ./hive --client-file ./configs/cancun.yaml --client go-ethereum,${cl}-bn,${cl}-vc --sim eth2/dencun" --sim.limit "eth2-deneb-sync/"

Run re-org tests

cl=lighthouse; ./hive --client-file ./configs/cancun.yaml --client go-ethereum,${cl}-bn,${cl}-vc --sim eth2/dencun" --sim.limit "eth2-deneb-reorg/"

[marioevz]

Changed back to draft because it requires changes in #898, and they might not even be included in master until we find a proper solution.

[marioevz]

I rebased and squashed most of the commits to be able to merge using a merge commit, so we preserve the multiple commits in the master branch.

I'm running locally tonight and should be able to merge tomorrow if all is good.