Closed andyleiserson closed 1 month ago
I think there should be a test that induces a DZKP validation failure and verifies (1) error from primary record, (2) error from non-primary records, (3) success of unaffected batch. But I didn't want to delay opening this PR to have that, and I may work on rebasing the batching changes first.
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 92.78%. Comparing base (
4fe0b41
) to head (d03c32f
).
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
Modify Fp25519 share conversion to support DZKP-based malicious security. By implementing the logic to periodically verify batches in the validator, significantly fewer changes to the protocol are required.
Additional changes are required to make batched proof verification work with seq_join -- right now, there is nothing to restrict seq_join (including validated_seq_join) from scheduling futures that should not be writing into the current batch.