proposer: prover-network proof-result caching #90
Description
Whenever the proposer is restarted during a GameProving task, or the task is retried due to an raising error condition (see #89),
it will submit the same proof inputs to the prover-network. This causes a lot of proof gas-spending and redundancy.
I can't think of a scenario where the same exact inputs to the proof would require a re-calculation of the proof, if it was successful.
Therefore we would require:
- a mapping of
input -> proof requestorinput-hash -> proof request - a mapping of
proof request -> (success, output)
Ideally this could be kept stateless and always fetched from the prover-network before a proof request would be submitted, but I don't know if this is possible.
A concrete feature request should be carefully determined prior.
Copy of https://github.com/celo-org/celo-blockchain-planning/issues/1308