谛听Diting Lab

Research Forums · Wed, Oct 7, 2026

Research discussions on ethresear.ch and Delving Bitcoin.

ethresear.ch

Who cleans up the code? SETCODEFROM's dangling bytecode and futureproof design choices

In briefWho cleans up the code? Dangling bytecode and design choices for SETCODEFROM

SummaryA discussion post on the Ethereum Research forum about bytecode that no account references anymore. It argues that the proposed SETCODEFROM and the partitioned binary tree turn this harmless leftover into a design decision. The post compares how clients handle the problem and why reference counting would be needed to delete shared code.

Delving Bitcoin

Anchoring an SPV light client on another chain: checkpoint trust, retargeting and reorg handling

In briefAnchoring an SPV light client on another chain: trust checkpoints, difficulty retargeting and reorgs

SummaryThe author describes the Bitcoin-facing design of Writz Protocol, which holds BTC collateral in a P2WSH script and runs an on-chain SPV light client on a smart contract platform. The post asks for critique of checkpoint trust, confirmation depth, timelock type, fee bumping, header validation cost and the reorg bound. The code is on testnet only and has not been audited.

ethresear.ch

Breaking the Serial Verification Loop via an Asynchronous Pipeline Consensus Architecture

In briefAsynchronous pipeline consensus to break the serial verification loop

SummaryThis forum post proposes an Asynchronous Pipeline Consensus Architecture that decouples block production from full transaction verification. Nodes propagate a compact payload of 6-byte transaction identifiers to start the next block, while verification of block n is delayed to block n+k via a 1-bit header field. The author asks whether the serial bottleneck can be broken without compromising baseline security and whether the design could apply to Ethereum's account model.