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
The new key getter from #7523 fails if no keys are registered and the canonical preimage is not known. We must however handle the scenario in which we can prove the keys were not distributed (if we don't happen to know them) to e.g. allow apps to skip encryption, note creation, etc., and so avoid King of the Hill issues (i.e. being prevented from creating a proof, leading to denial of service attacks).
A possible way would be to return Option<PublicKeys>, with Option::none meaning that keys are provably unkown. Currentl callsites would simply do .unwrap(), so there'd be little disruption.
The text was updated successfully, but these errors were encountered:
Key rotation is being dropped as of #8613, so we won't need to deal with that. The current scheme does allow for the preimage not being known, but we're migrating to a new one in which we'll be able to derive the ivpk from the address, so King of the Hill will not be an issue.
The new key getter from #7523 fails if no keys are registered and the canonical preimage is not known. We must however handle the scenario in which we can prove the keys were not distributed (if we don't happen to know them) to e.g. allow apps to skip encryption, note creation, etc., and so avoid King of the Hill issues (i.e. being prevented from creating a proof, leading to denial of service attacks).
A possible way would be to return
Option<PublicKeys>
, withOption::none
meaning that keys are provably unkown. Currentl callsites would simply do.unwrap()
, so there'd be little disruption.The text was updated successfully, but these errors were encountered: