Trace final actor state

[Stebalien]

At the moment, we trace the state of every invoked actor, but we don't include the state after it returns. For completeness, we should include the after state giving us:

1. The remaining balance.
2. The final state.
3. Whether or not it selfdestructed.

[rjan90]

@aarshkshah1992 can we sync on this issue this week. I can talk with Eva if this is really needed by Blockscout, etc.

[rvagg]

@Stebalien this is only with a tracing turned on, correct?

we trace the state of every invoked actor

This is with ExecutionEvent::InvokeActor, correct? So we'd introduce something like ExecutionEvent::ActorReturned which would include id, state and then a boolean for self-destruct? How would we determine that? Would we have to do a look-up on the actor to see if it still exists in the state at that point?

[Stebalien]

@Stebalien this is only with a tracing turned on, correct?

Yes.

This is with ExecutionEvent::InvokeActor, correct? So we'd introduce
something like ExecutionEvent::ActorReturned...

Yes. (probably ExecutionEvent::ExitActor, but yeah)

which would include id, state and then a boolean for self-destruct?

We shouldn't have to include the ID because it should always be paired with an InvokeActor.

How would we determine that? Would we have to do a look-up on the actor to see if it still exists in the state at that point?

Yes. Although that'll be cached so the check should be fast.