Added more tests for execution payload processing with incorrect blob commitments

[debjit-bw]

This PR expands the coverage of deneb's execution payload processing by adding the following two scenarios:

• test_no_commitments_for_transactions: while we test for incorrect commitments, this test adds a scenario where there are NO commitments even when blob txs are present. naturally the EL returns VALID, by definition. this test asserts that the CL doesn't reject it by running any validation.
• test_incorrect_commitments_for_no_blob: this test adds a scenario where none of the txs contain any blobs, but there are kzg commitments for them. it does so by combining the opaque txs of one sample tx with the commitments received from another. as earlier, the CL should not reject it.

[jtraglia]

Hmm this would be an invalid blob transaction though 😅

if get_tx_type(tx) == BLOB_TX_TYPE:
    # there must be at least one blob
    assert len(tx.blob_versioned_hashes) > 0

See: https://eips.ethereum.org/EIPS/eip-4844#execution-layer-validation

[debjit-bw]

oh no, this skipped my mind sorry, i removed this test

[jtraglia]

No worries at all. After looking at other tests in that file, it seems that some also have invalid transactions. So it might have been okay to keep but I think removing it is for the best. Thank you.

For example, an invalid transaction:

consensus-specs/tests/core/pyspec/eth2spec/test/deneb/block_processing/test_process_execution_payload.py

Lines 89 to 101 in b1c70a9

	def test_incorrect_transaction_length_1_extra_byte(spec, state):
	"""
	The versioned hashes are wrong, but the testing ExecutionEngine returns VALID by default.
	"""
	execution_payload = build_empty_execution_payload(spec, state)
	opaque_tx, _, blob_kzg_commitments, _ = get_sample_blob_tx(spec)
	opaque_tx = opaque_tx + b'\x12' # incorrect tx length, longer
	execution_payload.transactions = [opaque_tx]
	execution_payload.block_hash = compute_el_block_hash(spec, execution_payload, state)
	yield from run_execution_payload_processing(spec, state, execution_payload, blob_kzg_commitments)

[debjit-bw]

haha, hawk eyes 👀! btw, i just updated the base branch right after you approved the CI, would you mind approving the CI run once again?

[jtraglia]

Sure, we can add it back but with a new name. Something like this, feel free to alter:

- test_incorrect_commitments_for_no_blob
+ test_incorrect_transaction_no_blobs_but_with_commitments

Also I would add a comment (above the get_sample_blob_tx call) about blob transactions with zero blobs being invalid. This might not be immediately obvious to CL folks 😄

[jtraglia]

LGTM, thank you!