Marten Seemann

#1587 was merged, so I believe this should be fixed now. If not, please feel free to open a new issue.

> > It's probably easier if you use a slice of bufferEntrys. That slice could then start at length 1, and we'd append every minute, until we reach the maximum...

I think it can be a constant, the user chooses the duration (and that duration could, in principle, even be less than a minute).

> In the fraction, what represents token.IsRetryToken == false? From the name I think that it's a boolean that says if the user sent the correct retry token, but in...

> 1. Should we mock the two peers? We found `createPeer` function to instantiate the peers that need to connect with each other but we're not sure about the correctness...

> And perhaps mux is now github.com/libp2p/go-mplex, perhaps not??? Tidying deps would be great! No, it's not, and this is totally not what this issue is about. You just need...

> Hm. Isn't our early data a bit too early here? I assume the initiator would send early data on the second message (when the connection is encrypted, and the...

Side node: We might want to introduce a `CanListen` for transports, to avoid relying on the same assumption.

Rebased the PR. This should be ready for review now.

> I think this logic should change a bit after we discuss the multiaddr interpretation strategy. But at the very least I think we should be traversing the multiaddr backwards...