fix: handle report resubmit case

[Jeday]

Description

Scenario

• initial report is submitted at refSlot1
• next report is submitted at refSlot2, initial report is missed, warning event fired
  • WarnProcessingMissed(refSlot1)
• next report is re-agreed by consensus and resubmitted with new data
  • expected: WarnProcessingMissed(refSlot1)
  • actual: WarnProcessingMissed(refSlot2)

Solution

At resubmit we lack information about slot missed at prev submitting. So proposed solution is to prevent fire of WarnProcessingMissed event at resubmit. What will be achieved:

• Upholding 1 missed report - 1 event condition
• WarnProcessingMissed argument can be trusted to contain actual missed slot

Testing notes

Corresponding tests are located in the branch feature/shapella-upgrade-tests-base-oracle

[ujenjt]

looks valid, @skozin could you check plz

[ujenjt]

also, we don't have an event for slot re-agreement occurrences. Maybe let's add one for monitoring purposes?

[TheDZhon]

I think the fix should be applied, though waiting for Sam's feedback

[Jeday]

also, we don't have an event for slot re-agreement occurrences. Maybe let's add one for monitoring purposes?

This would be more relevant in HashConsensus contract. I think it can be implemented fairly easily. Though this brings us to question: if consensus can be reached again with different result and report submitted, why can't report be withdrawn if consensus is lost?

[skozin]

also, we don't have an event for slot re-agreement occurrences. Maybe let's add one for monitoring purposes?

good idea, will add in a separate PR

[skozin]

I'd say that firing WarnProcessingMissed(refSlot1) twice is not the expected behavior, it's also incorrect. And the expected one is not firing the event twice, exactly the one this PR implements.