EIP-8304: Trustless log and transaction index

⌬LLM No new opcode is defined.

⚇Human No new opcodes.

⌬LLM No existing opcode result or non-gas behavior is modified or deprecated.

⚇Human LOG0–LOG4 and all other opcodes retain their existing execution semantics.

⌬LLM No precompile is introduced.

⚇Human No new precompiles.

⌬LLM No precompile logic or gas schedule is modified.

⚇Human No precompiles are modified.

⌬LLM Exactly one new system contract is added, and it is stateful because mandatory block processing repeatedly writes its storage.

⚇Human One new stateful system contract is introduced. It receives a protocol-generated update every block and may receive multiple updates when several table levels become publishable at the same block.

⌬LLM Reusing SYSTEM_ADDRESS and the system-call convention does not modify the code, state, or behavior of the pre-existing beacon-roots system contract.

⚇Human No pre-existing system contract is modified.

⌬LLM The new work is a block-level post-transaction operation; no opcode's internal state-access point or gas-charge ordering is changed.

⚇Human No changes to State-access ordering within opcode execution

⌬LLM No blob gas rule, charge, budget, or field is introduced or modified.

⚇Human Blob gas is not involved.

⌬LLM The stateful contract adds storage writes but defines no new state-gas rate, charging site, budget, reservoir, or execution-gas spill rule.

⚇Human No state gas cost changes.

⌬LLM No gas-refund mechanism is added or changed.

⚇Human Refunds are not involved.

⌬LLM Indexing transactions does not introduce a transaction envelope or type.

⚇Human No new transaction type.

⌬LLM No existing transaction's validity rules or intrinsic gas calculation changes; the proposal derives a post-execution commitment from already processed transactions.

⚇Human Transaction validity, intrinsic gas, and transaction encoding are all unchanged.

⌬LLM No block or header field is added; the new commitment is part of execution state.

⚇Human No header field is added;

⌬LLM The proposal defines a new internal index-table encoding but does not change transaction, block, header, or Engine-interface encoding. The anchor's enumerated protocol encoding surfaces therefore remain unchanged.

⚇Human The proposal defines an internal encoding for index entries and uses SSZ for table commitments, but it does not alter transaction, receipt, block, header, or protocol-interface encoding.

⌬LLM Regeneration adds local post-sync work, but no block-RLP validation mechanism or sync-protocol encoding is introduced, which is the scope of this anchor.

⚇Human No block RLP validation rule or synchronization protocol is modified.

⌬LLM Activation installs code and initializes an account nonce, an explicit irregular state modification at the fork boundary, and begins table scheduling from that boundary.

⚇Human Fork activation requires deploying/initializing the new system-contract account. In addition, a table is only generated when the EIP was already active at its `first_block`, so each table level starts appearing progressively after the fork. Explicit fork-boundary state-transition tests are required.

⌬LLM No Engine API field, endpoint, or communication mechanism is specified.

⚇Human No execution-payload field, Engine API endpoint changes.

⌬LLM A single-block transition interface needs a new mechanism to supply or retain recent blocks and prior table material so it can derive periodic roots. That is more than one isolated scalar field, but the EIP does not prescribe the exact interface.

⚇Human The transition tool needs a new index-table construction mechanism. A level-0 table can be derived from the current block and its receipts, but publishing delayed higher-level tables requires prior index entries, historical blocks, or equivalent persistent context that is not available in an ordinary single-block transition input.

⌬LLM Root correctness and update availability affect trustless proofs and consensus state, while hot storage creates a targeted resource-attack surface. The interactions are concentrated in index derivation and its system contract, so targeted review and fuzzing fit better than the broadest security anchor.

⚇Human Incorrect entry generation, ordering, SSZ root, delay timing, or ring-buffer slot all produce a different state root and therefore consensus divergence.

⌬LLM Work scales with transaction and log content, interacts with receipt production, periodic multi-level merging, block deadlines, storage updates, and sync recovery. End-to-end and adversarial-content performance cannot be established by one isolated microbenchmark and can substantially affect validation behavior.

⚇Human Every block must build, encode, sort, and hash all transaction and log entries, and merge the 4/16/64/256-block tables before their deadlines. The cost interacts with tx count, log count, topics, reorgs, disk layout, and sync state, and cannot be fully validated by isolated microbenchmarks alone.

⌬LLM Genesis, activation, empty tables, level alignment, delayed availability, five table sizes, and 1,024-slot wraparound create multiple interacting boundary mechanisms. The schedule-by-level and overwrite combinations require an elevated case count.

⚇Human Covers the fork boundary, genesis, the parent-block hash offset, table alignment, five table sizes, delay deadlines, ring-buffer wraparound, invalid/zero table size, zero root, empty blocks, topic count, duplicate values, lexicographic byte ordering, reorgs, and sync initialization. A large combinatorial test set is required.

⌬LLM The proposal introduces one commitment construction composed of established SHA2-256 hashing and SSZ Merkleization; it is well known rather than novel.

⚇Human Adds SHA-256 entry hashing and SSZ merkleization, but both are mature mechanisms with abundant existing test resources.