E2E tests for indexer fallbacks (EIP4844, calldata)

[nuke-web3]

WIP

followup to #530

• exposes and documents setting beacon RPC endpoint (needed for EIP4844 blobs)
• {more to come}

[nuke-web3]

Working MVP! Tested against https://github.com/nuke-web3/ethereum-docker/tree/n/add-anvil-and-op setup building bin & running for node, proposer, and batcher (see ./scripts in the repo)

indexer works to check calldata and blobs.

TODO

• confirm celestia works and indexer responds correctly (need docs on setup)
• update readme with indexer response example for Celestia
• document how to use & test manually
• consider adding integration e2e test against op devnet
• fix existing e2e failing test

---

Interesting Finding:

When switching on a live chain to/from blobs and calldata, the op-proposer seems to re-batch and publish all L2 blocks and the indexer (correctly I think) overwrites where L2 block data is located. It is unclear to me if this is the behavior intended or operator error 😅

# op-batcher
# https://docs.optimism.io/chain-operators/guides/configuration/batcher#batcher-configuration
# [calldata|blobs|auto] The default value is calldata.
OP_BATCHER_DATA_AVAILABILITY_TYPE=blobs

[tuxcanfly]

OK.