File sharing example cannot support the sample plot situation

[YoshieraHuang]

Hi, I am new to libp2p but very interested for this. While surfing through examples, I find the provided file-sharing example are not so ideal. I run these three scripts at the base directory of the repository in different console to test the sample plot.

In first console:

 cargo run --example file-sharing -- --listen-address /ip4/127.0.0.1/tcp/40837 --secret-key-seed 1 provide --path ./Cargo.lock --name lock

for serving Cargo.lock file with name lock. It works. The listening multiaddr is /ip4/127.0.0.1/tcp/40837/p2p/12D3KooWPjceQrSwdWXPyLLeABRXmuqt69Rg3sBYbU1Nft9HyQ6X

In second console:

cargo run --example file-sharing -- --listen-address /ip4/127.0.0.1/tcp/40800 --peer /ip4/127.0.0.1/tcp/40837/p2p/12D3KooWPjceQrSwdWXPyLLeABRXmuqt69Rg3sBYbU1Nft9HyQ6X --secret-key-seed 2 provide --path ./Cargo.toml --name toml

for serving Cargo.toml file with name toml and connect to the first console. It also works. The listening multiaddr is /ip4/127.0.0.1/tcp/40800/p2p/12D3KooWH3uVF6wv47WnArKHk5p6cvgCJEb74UTmxztmQDc298L3.

In thrid console, I connect to the second console and request file lock in first console:

cargo run --example file-sharing -- --peer /ip4/127.0.0.1/tcp/40800/p2p/12D3KooWH3uVF6wv47WnArKHk5p6cvgCJEb74UTmxztmQDc298L3 get --name toml

Now the third console panics with this informations:

thread 'async-std/runtime' panicked at 'Dialing(PeerId("12D3KooWPjceQrSwdWXPyLLeABRXmuqt69Rg3sBYbU1Nft9HyQ6X"))', examples/file-sharing.rs:502:22
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
thread 'main' panicked at 'Sender not to be dropped.: Canceled', examples/file-sharing.rs:332:28

Meanwhile, the first console also panics with this informations:

thread 'async-std/runtime' panicked at 'IncomingConnectionError { local_addr: "/ip4/127.0.0.1/tcp/40837", send_back_addr: "/ip4/127.0.0.1/tcp/52526", error: Transport(Other(Custom { kind: Other, error: Other(A(B(Select(Failed)))) })) }', examples/file-sharing.rs:502:22
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
thread 'main' panicked at 'not yet implemented', examples/file-sharing.rs:141:26

It seems the connection from the third console to the first console fails.

[YoshieraHuang]

Requesting the toml file served in the second console from the third console in this situation also fail.

[mxinden]

Interesting, mind printing the event that is not expected via the change below?

diff --git a/examples/file-sharing.rs b/examples/file-sharing.rs
index 183bd9d6..2c069175 100644
--- a/examples/file-sharing.rs
+++ b/examples/file-sharing.rs
@@ -138,7 +138,7 @@ async fn main() -> Result<(), Box<dyn Error>> {
                             network_client.respond_file(file_content, channel).await;
                         }
                     }
-                    _ => todo!(),
+                    e => todo!("{:?}", e),
                 }
             }
         }

[YoshieraHuang]

Interesting, mind printing the event that is not expected via the change below?

diff --git a/examples/file-sharing.rs b/examples/file-sharing.rs
index 183bd9d6..2c069175 100644
--- a/examples/file-sharing.rs
+++ b/examples/file-sharing.rs
@@ -138,7 +138,7 @@ async fn main() -> Result<(), Box<dyn Error>> {
                             network_client.respond_file(file_content, channel).await;
                         }
                     }
-                    _ => todo!(),
+                    e => todo!("{:?}", e),
                 }
             }
         }

No unexpected network event happens. The file server and the request client both panic during handling swarm event: https://github.com/libp2p/rust-libp2p/blob/a168410dbed0d0941f2e5a14543206044ccb2260/examples/file-sharing.rs#L502

[mxinden]

And on which SwarmEvent are they panicing? When you exclude that event from panic do things function as expected?

[YoshieraHuang]

And on which SwarmEvent are they panicing? When you exclude that event from panic do things function as expected?

I excluded the IncomingConnectionError and Dialing SwarmEvent from panic and the file-sharing example works as expected. But I wonder why these two events emit. It seems that the client try to connect to the peer which is not directly connected but it fails.

[mxinden]

IncomingConnectionError might be that the node is failing to connect to a private IP address?

I excluded the IncomingConnectionError and Dialing SwarmEvent from panic and the file-sharing example works as expected.

Mind creating a pull request?

[YoshieraHuang]

After the error is solved superficially, I still wonder why the client will dail to a unconnected server. It's a behavior of kadmelia protocol?

[mxinden]

Nodes will try to interconnect due to the Kademlia protocol, yes.

[thomaseizinger]

Let me know if still an issue.