feat(gossipsub): implement gossipsub 1.2 beta

[hopinheimer]

Description

This PR implements gossipsub 1.2 beta bringing changes over from lighthouse
ref PR: sigp/lighthouse#5422

Please include any relevant issues in here, for example:
libp2p/specs#548

[mergify]

This pull request has merge conflicts. Could you please resolve them @hopinheimer? 🙏

[elenaf9]

Thanks @hopinheimer!

Couple of comments.
I think the new protocol version also needs to be added to the ProtocolConfig?

[hopinheimer]

Hi @elenaf9 thank you for the review, I have made all the requested changes although, there seem to be a test that's failing test_slow_peer_returns_failed_ihave_handling() I debugged and found the events were null. could you guide me to what might be the issue?

[elenaf9]

Hi @elenaf9 thank you for the review, I have made all the requested changes although, there seem to be a test that's failing test_slow_peer_returns_failed_ihave_handling() I debugged and found the events were null. could you guide me to what might be the issue?

I think it has to do something with gossip_promises now not depending on peer_score being activated anymore. Thus, the gossipsub_promises logic is now also there even when peer_score hasn't been activated.

handle_ihave checks if the message id is already in the pending promises, and doesn't send another message to the peer if it is. Thus in the test the second call to gs.handle_ihave is a no-op. Given that the sender to the slow peer's connection handler has capacity of 2, it therefore won't be full yet and not report a SlowPeer event.
Before, this issue didn't come up I think because peer_score (and thus gossip_promises) isn't enabled it the test.

The test can be fixed by using a different MessageId in the second handle_ihave call.

---

I was actually wondering if it was on purpose that gossip_promises is now always active and doesn't depend on peer_score anymore. @jxs as the author of sigp/lighthouse#5422, could you briefly explain why this decision was made?