[test scenario] windowed PoSt - behaviour under chain reorgs

[raulk]

Describe the test scenario.

This is part of the Windowed PoSt epic. Imagine that you’re a miner mining on chain A, and you posted a PoSt correctly under the right window, 4 epochs before it was about to close. However, there was a fork in the network, and chain B appears with a heavier weight which anchors to H-5. Everyone, including you, switches to chain B, therefore your proof is lost.

First:

• [ ] confirm what should happen here (message posted on Lotus slack).
• [ ] verify that indeed that's what happens, repeatedly.

Provide any background and technical implementation details.

What should we measure?

Discomfort factor (0-10).

High (8-10).

Additional remarks.

[raulk]

Low discomfort factor suggests we should deprioritise this one and remove it from pre-mainnet sprints.