You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Fendermint currently doesn't have a concept of delegated addresses. Forest has an is_delegated method but at the time when I looked it didn't use it for anything, so I didn't investigate and only used the SignatureType from fvm_shared which doesn't have the concept of a delegated type, so I did the same check as Forest, which includes no domain separation (ie. no chain ID). Forest uses a shim for the signature, which is where they added this new type, no doubt for future use (I even saw the Slack convo where they asked why it's not in the FVM).
As I understand FIP-55 has to be implemented by the node for FEVM type messages, and to protect against replay attacks across subnets, it would be enough to hash the chain ID (which the node knows) together with the CID of the message, without adding it to the message. A new --chain-id option has to be added to all the CLI commands, though, and the MessageFactory in the RPC library.
The text was updated successfully, but these errors were encountered:
Originated from this Slack discussion.
This is the FIP explaining how the FEVM chain ID separation works: https://github.com/filecoin-project/FIPs/blob/3a4ba11c41e8a8e641d27b8999dab3a774b2751e/FIPS/fip-0055.md#delegated-signature-type
And this one never got implemented: https://github.com/filecoin-project/FIPs/blob/master/FIPS/fip-0039.md
Fendermint currently doesn't have a concept of delegated addresses. Forest has an is_delegated method but at the time when I looked it didn't use it for anything, so I didn't investigate and only used the SignatureType from
fvm_shared
which doesn't have the concept of a delegated type, so I did the same check as Forest, which includes no domain separation (ie. no chain ID). Forest uses a shim for the signature, which is where they added this new type, no doubt for future use (I even saw the Slack convo where they asked why it's not in the FVM).As I understand FIP-55 has to be implemented by the node for FEVM type messages, and to protect against replay attacks across subnets, it would be enough to hash the chain ID (which the node knows) together with the CID of the message, without adding it to the message. A new
--chain-id
option has to be added to all the CLI commands, though, and theMessageFactory
in the RPC library.The text was updated successfully, but these errors were encountered: