Thomas Eizinger

I agree with your concerns. One thing to note is that the extra match will ident all code handling the event by two tabs: one for the `match` and one...

If we go with an enum, we could probably also unify `SwarmEvent` and the above `InEvent`.

Thinking about it more, a single `on_event` function seems like the right primitive to me. Everything else, including the macro above, can be built on top of that.

One problem we currently have is that with `NotifyHandler` on `NetworkBehaviourAction`, we cannot notify all connections which makes it cumbersome to use the `InEvent::UpdateState` variant. @mxinden @elenaf9 What do you...

I looked at using this implementation for the ping protocol and I think it would be doable apart from I thing: Overriding `keep_alive`. The current ping `ConnectionHandler` allows the user...

> I have to give this abstraction more thought. I find this abstraction to be very similar to `libp2p-request-response`. I have to think about whether it makes sense to have...

> > @mxinden @elenaf9 What do you think of adding `NotifyHandler::All`? > > I don't see a reason not to add it, and updating the state of all handlers sounds...

> > I tried using this locally for rendezvous and we would have to use this mechanism to push the registration data from the behaviour down to the connections. >...

> > > > @mxinden @elenaf9 What do you think of adding `NotifyHandler::All`? > > > > > > > > > I don't see a reason not to add...

@mxinden @elenaf9 Please have another look at this! I also added a dirty patch that converts `libp2p-rendezous` to use the new `FromFn` handler. I also added a `Shared` abstraction that...