early rewards #233
early rewards #233
Conversation
pifragile
commented
Jun 17, 2022
pifragile
commented
- Coses [rpc] verify meetup already during attesting phase #178
| legit_participants: Participants, | ||
| participant_attestations: Attestations, | ||
| ) -> bool { | ||
| for (i, mut attestations) in participant_attestations.into_iter().enumerate() { |
There was a problem hiding this comment.
what happens, if i attested people that fell out of the legit participant set? I believe your logic with sorting and testing equality would not see this as fully connected. but it should
Anyway, sorting twice within a for loop doesn't strike me as efficient. Wouldn't a nested for loop through legit participants be more efficient?
There was a problem hiding this comment.
this test is made before any participants can fall out of the set, because the attestation graph needs to be fully connected for all participants in order for early payout to be possible, right?
about the efficiency. my implementation:
O(n) (loop over attestations) * O(n^2) (sorting, worstcase)
nested loops:
O(n) (loop over attestations) * O(n) (loop over each elem in attestations) * O(n) (check if legit_participant contains elem)
so in asymptotic complexity they are the same.
i agree that in the first implementation we have some overhead because it is 2*O(n^2), but i would argue that for those small input sizes it shouldnt really matter.
but you decide, should i change it to the nested loop?
There was a problem hiding this comment.
nah, its bounded, so I guess it's fine. Do we already have a worst case benchmark for this claim function? that will tell us the price
|
I need more time to review the tests, do not merge yet |
