Responding to the 7702 authorizations that were stored and replayed in the Binance incident, EIP-8477 proposes binding every authorization to one of the last 256 blocks so that it expires. Frame Transaction got two revisions: lightclient added lean SPHINCS post-quantum signatures, and nerolation proposed leaving the calldata cost of signatures out of MAX_VERIFY_GAS. A batch of non-author PRs corrected the wording of EIP-2780, EIP-8037, EIP-8038 and other Glamsterdam specs. L2BEAT was renamed ETHBEAT.
Opened at 00:39 by dangerousfood, names the 7702 sweeps in the Binance incident as its motivation. In addition to chain ID, delegate address and nonce, every authorization also signs a recent execution block number and block header hash. When included in block N, the signed block must be one of the 256 ancestors before N with a matching hash. Exactly 256 blocks back is still valid, 257 is not, and neither the current block nor future blocks count. Freshness is checked before signature recovery, and stale or mismatched authorizations are simply skipped.
Opened at 15:59 by nerolation. Mempool rule 6 originally said "the intrinsic cost of verifying `tx.signatures`", which could be read as including the calldata cost of the signature bytes. The revision changes it to `signature_verification_cost` and states that `frame_data_cost` and `signature_data_cost` are not counted. The reasoning: MAX_VERIFY_GAS limits the validation work a node does, while calldata cost pays for bandwidth and block space. If calldata were counted, a PQ signature would exceed the 100,000 limit on size alone.
At 21:26 flcl42 submitted 7 non-author revisions at once: EIPs PR #12622: EIP-2780 fixes the floor reference and specifies delegation target warming (in Glamsterdam the floor uses EIP-7976's `floor_tokens_in_calldata`, 4 tokens per byte and 16 gas per token, plus the EIP-7981 access list data surcharge; execution-specs already follows the corrected text), EIPs PR #12624: EIP-8037 fixes wording that contradicts the spec, EIPs PR #12626: EIP-8038 scopes the ACCOUNT_WRITE rule and the worst-case client claim, EIPs PR #12623: EIP-7708 fixes gas numbers and creation transaction rules, EIPs PR #12625: EIP-7778 removes a contradictory restore-original refund example, plus one each for EIP-7928 and EIP-7843.
At 21:26 flcl42 submitted 7 non-author revisions at once: EIPs PR #12622: EIP-2780 fixes the floor reference and specifies delegation target warming (in Glamsterdam the floor uses EIP-7976's `floor_tokens_in_calldata`, 4 tokens per byte and 16 gas per token, plus the EIP-7981 access list data surcharge; execution-specs already follows the corrected text), EIPs PR #12624: EIP-8037 fixes wording that contradicts the spec, EIPs PR #12626: EIP-8038 scopes the ACCOUNT_WRITE rule and the worst-case client claim, EIPs PR #12623: EIP-7708 fixes gas numbers and creation transaction rules, EIPs PR #12625: EIP-7778 removes a contradictory restore-original refund example, plus one each for EIP-7928 and EIP-7843.
At 21:26 flcl42 submitted 7 non-author revisions at once: EIPs PR #12622: EIP-2780 fixes the floor reference and specifies delegation target warming (in Glamsterdam the floor uses EIP-7976's `floor_tokens_in_calldata`, 4 tokens per byte and 16 gas per token, plus the EIP-7981 access list data surcharge; execution-specs already follows the corrected text), EIPs PR #12624: EIP-8037 fixes wording that contradicts the spec, EIPs PR #12626: EIP-8038 scopes the ACCOUNT_WRITE rule and the worst-case client claim, EIPs PR #12623: EIP-7708 fixes gas numbers and creation transaction rules, EIPs PR #12625: EIP-7778 removes a contradictory restore-original refund example, plus one each for EIP-7928 and EIP-7843.
At 21:26 flcl42 submitted 7 non-author revisions at once: EIPs PR #12622: EIP-2780 fixes the floor reference and specifies delegation target warming (in Glamsterdam the floor uses EIP-7976's `floor_tokens_in_calldata`, 4 tokens per byte and 16 gas per token, plus the EIP-7981 access list data surcharge; execution-specs already follows the corrected text), EIPs PR #12624: EIP-8037 fixes wording that contradicts the spec, EIPs PR #12626: EIP-8038 scopes the ACCOUNT_WRITE rule and the worst-case client claim, EIPs PR #12623: EIP-7708 fixes gas numbers and creation transaction rules, EIPs PR #12625: EIP-7778 removes a contradictory restore-original refund example, plus one each for EIP-7928 and EIP-7843.
At 21:26 flcl42 submitted 7 non-author revisions at once: EIPs PR #12622: EIP-2780 fixes the floor reference and specifies delegation target warming (in Glamsterdam the floor uses EIP-7976's `floor_tokens_in_calldata`, 4 tokens per byte and 16 gas per token, plus the EIP-7981 access list data surcharge; execution-specs already follows the corrected text), EIPs PR #12624: EIP-8037 fixes wording that contradicts the spec, EIPs PR #12626: EIP-8038 scopes the ACCOUNT_WRITE rule and the worst-case client claim, EIPs PR #12623: EIP-7708 fixes gas numbers and creation transaction rules, EIPs PR #12625: EIP-7778 removes a contradictory restore-original refund example, plus one each for EIP-7928 and EIP-7843.
EIPs PR #12619: Add EIP-8484 SLOTDURATION Opcode, opened at 20:37 (Magicians: EIP-8484 SLOTDURATION opcode discussion thread), preparing for contracts to read the slot duration once slots get shorter.
13:28; Barnabé: the demos include filling a CoW order with L1 liquidity, changing a Safe owner with L1 as the root of trust, and authorizing an L1 transfer from a rollup, 15:58.
EEZ: three first demos of synchronous composability between rollups and mainnet, 13:28; Barnabé: the demos include filling a CoW order with L1 liquidity, changing a Safe owner with L1 as the root of trust, and authorizing an L1 transfer from a rollup, 15:58.
00:22, asking "where is the accountability"; Fede's reply, 21:54, argues Ethereum's problems are mostly cultural and that performance and UX have been improving in recent years; binji's long reply, 23:18.
Georgios Konstantopoulos criticizes the Devcon keynotes, 00:22, asking "where is the accountability"; Fede's reply, 21:54, argues Ethereum's problems are mostly cultural and that performance and UX have been improving in recent years; binji's long reply, 23:18.
Georgios Konstantopoulos criticizes the Devcon keynotes, 00:22, asking "where is the accountability"; Fede's reply, 21:54, argues Ethereum's problems are mostly cultural and that performance and UX have been improving in recent years; binji's long reply, 23:18.
