vyzo

> So sometimes I can see the multiaddr but sometimes not. Why? not sure about that. > Why was there a host with port 59919 created for 5000? that's an...

Sure, that would be very useful. Care to take a stub at it? We can guide you through.

With signatures we can easily cache validation results (ie have a second time cache that tracks `valid` messages) and deliver duplicates without re-entering the validation pipeline.

The key should be just the message Id, no need for anything else. And yes, it would be a timecache with some appropriate duration. But don't rush to implement this...

We would still have to validate the signature, but in general we use messageID's for tracking things internally. A peer sending us a message with the same seqNo and different...

Drop the message and `Warn` is probably fine; but we might want to consider blacklisting in this case as well. I am just a little weary of the pubsub library...

Well, it's not a correctness problem, but indeed a malicious peer could cause us to allocate an unlimited number of streams so it is a bug. We can add a...

I still think we actually _can_ do it. There is nothing in the protocol that requires more than two streams and we use a single stream created at connection time....

I meant looking at the number of streams within the connection. The new connection shouldn't be affected by the old streams.

I think that #332 is kind of hard to implement without breaking changes anyway.