Another problem there is that transaction types may have different meaning across chains. Only the legacy transaction type (without chain ID) is truly replayable across networks, for the rest there is no guarantee.
Can you give a few examples in elaborating what you mean?
Hey, may I ask what the current status of this EIP is? As I understand it’s still in review phase? I think this direcetion is very bad, we really need collision resistance, and having an onchain registry for chains.
Another motivation: https://twitter.com/koeppelmann
ENSIP-11 to be included in reference
@bumblefudge and @ligi ,
I am thinking of changing this EIP’s track from a Standard Core EIP (which is enforcible and requires client update) to a Informational EIP that outlines a list of recommended ranges of ChainID and their implications but no enforcement nor client codechange.
I think as long as ppl are aware and use them that way and all dApps and block explorers and wallets assumes it those ways, it should be good enough, I guess?
Per discussion with @bumblefudge and @pedrouid, I am reviving EIP-2294 draft and modify it to enable this CAIP: Create caip-draft-evm-xaddress.md by xinbenlv · Pull Request #304 · ChainAgnostic/CAIPs · GitHub
ChainId=0
seems to be getting special semantics in EIP-7702 where authorization signatures with ChainId=0
are deemed to be chain agnostic.