Double voting

[dnkolegov]

The issue is described here - consensus-shipyard/ipc-actors#111

[mikirov]

With our current design we track the votes of a validator for a given epoch. The problem can be solved by removing the validator votes for the current back to the last executed epoch when they leave so their votes don't persist after they're gone. This solution however requires us to for loop over the previous epochs, which is not optimal in blockchain. Do we still want to go with this approach?

[mikirov]

If the validator leaves we also have the problem that a different submission than the one he voted for may reach majority. So the whole executable submission logic would need to be re-evaluated on validator leave. This requires fundamental change in the contract logic.

[mikirov]

What is your take on the isssue @adlrocha ?

[adlrocha]

In the new implementation of the contracts validators no longer need to vote for checkpoints. @dnkolegov, feel free to reopen if needed.