Context

That's the continuation of %DAL: baker incentives (part 3: protocol ready) and %DAL: baker incentives (part 4: accuser). The goal is to test the DAL incentives scheme in order to eliminate possible bugs.

Participants

@eugenz (owner), @phink, @iguerNL

Work breakdown

Tests are listed more or less according to their priority, and are separated in whether they target the protocol code or not.

For the protocol code:

• (@eugenz, ETA: 13/02) integration/unit tests for the accusation operation
  • valid accusations can be injected (!16277 (merged))
  • check that each validation error is triggered as expected (!16375 (merged) and !16693 (merged))
  • check that the misbehaving baker gets no DAL rewards (also checked in !16515 (merged))
  • check scenario from issue #7704 (closed) (!16693 (merged))
  • check that there is enough slot header history to check accusations #7653 (closed) (!16693 (merged))
• (@iguerNL, ETA: 30/01, !16355 (merged)) integration tests for rewards (#7602 (closed))
  • full rewards when participation is met (!16355 (merged))
  • no rewards when participation is not met (!16355 (merged))
  • full rewards if no DAL activity (!16506 (merged))
  • reward amount as expected (!16355 (merged))
• (@eugenz, ETA: 13/02) integration tests for rewards
  • check output of dal_participation (#7682 (closed), !16506 (merged))
  • add a regression test for dal_participation RPC in tezt/tests/RPC_test.ml (#7697 (closed), !16506 (merged))
  • check that expected_dal_rewards = expected_assigned_shards_per_slot * dal_attesting_reward_per_shard where the values are taken from dal_participation and expected_issuance (!16506 (merged))
  • add regression tests for DAL rewards (#7711 (closed), !16506 (merged))
  • check rewards for a small baker (which doesn't have shards at each level) (!16809 (merged))
• (@eugenz, ETA: 04/02) mainnet migration test

Non-protocol tests:

• (@phink, ETA: 30/01) integration test for the DAL node accuser
  • integration test using the DAL node accuser (!16515 (merged))
    • be able to run a modified attester node on the CI (!16515 (merged))
  • different scenarios for accusations' injections (!16389 (merged))
• (@eugenz, ETA: 17/02) test share_is_trap (!17009 (merged))
  • check that the number of traps is as expected over many shards; use eg 1000 times more shards than what is needed for one trap for traps_fraction = 0.0005, that’s 2000 * 1000 shards; approach: generate many slots, use some real phks
  • edge cases: traps_fraction is 0 or 1
  • check the running time (just to check there are no surprises)