Removes the usage of the `to_swarm` `libp2p-swarm-derive` attribute in favor of the automatically generated event through the `NetworkBehaviour` derive macro.
Pull-Request: #4580.
When a gateway is not found or is not routable, UPnP `NetworkBehaviour`'s `poll` repeatedly returns the same failure state, which increases CPU usage.
Pull-Request: #4569.
According to the [spec], all fields within push messages are optional. To handle missing fields, we deserialize push messages into a different struct, patch the previously received, full identify message with it and emit this as the new info.
Previously, we failed to parse the message which causes incompatibility with js-libp2p nodes as they don't send the public key in push messages. See 88c47f51f9/packages/libp2p/src/identify/identify.ts (L205-L210).
[spec]: https://github.com/libp2p/specs/tree/master/identify#identifypush
Pull-Request: #4495.
Renamed the following
- `dns::GenDnsConfig` -> `dns::Config`
- `dns::DnsConfig` -> `dns::async_std::Config`
- `dns::TokioDnsConfig` -> `dns::tokio::Config`
If async-std feature is enable, use `dns::async_std::Config`. When using tokio, import `dns::tokio::Config` . There is no need to use `dns::Config` directly.
Resolves#4486.
Related: #2217.
Pull-Request: #4505.
This is a relict from when we still had to delay the construction of `ToSwarm` commands because some data was only available in `poll` via `PollParameters`. This is now resolved and `PollParameters` will go away.
Pull-Request: #4536.
Previously, a connection would be shut down immediately as soon as its `ConnectionHandler` reports `KeepAlive::No`. As we have gained experience with libp2p, it turned out that this isn't ideal.
For one, tests often need to keep connections alive longer than the configured protocols require. Plus, some usecases require connections to be kept alive in general.
Both of these needs are currently served by the `keep_alive::Behaviour`. That one does essentially nothing other than statically returning `KeepAlive::Yes` from its `ConnectionHandler`.
It makes much more sense to deprecate `keep_alive::Behaviour` and instead allow users to globally configure an `idle_conncetion_timeout` on the `Swarm`. This timeout comes into effect once a `ConnectionHandler` reports `KeepAlive::No`. To start with, this timeout is 0. Together with https://github.com/libp2p/rust-libp2p/issues/3844, this will allow us to move towards a much more aggressive closing of idle connections, together with a more ergonomic way of opting out of this behaviour.
Fixes#4121.
Pull-Request: #4161.
Implements UPnP via the IGD protocol. The usage of IGD is an implementation detail and is planned to be extended to support NATpnp.
Resolves: #3903.
Pull-Request: #4156.
This PR adds a delay to ping connection handler after exceeding the failure rate to prevent opening a outbound stream right after an error.
Resolves#4410.
Pull-Request: #4423.
This PR changes the logging of the Kademlia connection handler related to the remote Kademlia mode changes:
- Downgrade log level for the remote kademlia protocol report from `info` to `debug`.
- Introduce connection_id for the handler to improve logging.
Pull-Request: #4278.
Added a getter function to obtain `TopicScoreParams`. I need this function for testing which checks if scoring parameters are updated as expected.
Pull-Request: #4231.
`libp2p-relay` crate specifies the `libp2p-swarm/async-std` feature, which causes an `async-std` dependency to always be introduced.
Pull-Request: #4283.
A few weeks ago when I was debugging my wrong setup with kademlia, I could not conveniently debug `RoutingUpdate` after adding an address to the DHT. It would be nice to have a few derivable traits on a public enum.
Pull-Request: #4270.
The `rate-limiter` module was not made publicly available. This change exports the `RateLimiter` trait defined within that module which allows users to define their own rate limiters. To make common usecases easy, we also provide a set of constructor functions for common rate limiters.
Resolves#3741.
Pull-Request: #3742.
By default, dialing is prevented when already connected. `kad` already checks this manually before doing dials. By changing the condition from `Disconnected` to `NotDialing`, we prevent simultaneous dials to a peer.
Pull-Request: #4224.
