moshe-blox
commented
4 months ago Good!
The only thing we have to think about before merging this PR is that for committee runner the code was initially designed to work if a post-consensus was missing a root (in case there was an issue with a certain validator).
Not sure if we need to really care for this case though... Let's talk about it tomorrow
Yes, either missing or has extra roots, because currently SSV doesn't guarantee equal validator sets between operators at any given time.
If a validator registered with 1 corrupt BLS share, only 1 operator would know it and they would send post consensus messages with 1 missing signature to the other operators, and the opposite for that operator which would see 1 extra signature from others.
The other reason for mismatching validator sets is contract event syncing disparity.
This should be solved in the future, but its not part of the plan for Committee Consensus.
Maybe we can just check that an operator did not send more than some fixed reasonable limit of signatures and that they are all with unique validator indices?
GalRogozinski
MatheusFranco99
This PR adds a validation on the number of signing roots (and their content) for the CommitteeRunner and aligns the existing tests with the new error.