feat: add peerinfo by varex83 · Pull Request #103 · NethermindEth/pluto

varex83 · 2026-01-05T13:29:10Z

Uses libp2p::ping as a reference https://github.com/libp2p/rust-libp2p/blob/4ee0e07dc01b5190026425d7e28872fefe0e6f3c/protocols/ping/src/handler.rs

github-actions · 2026-01-05T13:41:47Z

Coverage (base → head): 35.44% → 34.54% ⬇️ 0.9 pp
Only posts when PR coverage is lower than the base branch's latest push (or freshly computed base).

emlautarom1

For this PR in general is a matter of trust: I cannot really validate that the implementation is correct and I would not be able to debug if there were issues. Also, the structure is completely different to the Go implementation.

We can go with this approach but I wonder if relying on libp2p_stream wouldn't be easier to debug and result in less code to maintain. I would like to have a second opinion about it (cc. @iamquang95).

emlautarom1 · 2026-01-12T21:01:15Z

+futures = "0.3"
+futures-timer = "3.0"


Dependency versions should be listed in the top level Cargo.toml file only.

my bad, forgot to move

emlautarom1 · 2026-01-12T21:11:41Z

+    // Encode message to protobuf bytes
+    let mut buf = Vec::with_capacity(msg.encoded_len());
+    msg.encode(&mut buf)
+        .map_err(|e| io::Error::new(io::ErrorKind::InvalidData, e))?;
+
+    // Write unsigned varint length prefix
+    let mut len_buf = unsigned_varint::encode::usize_buffer();
+    let encoded_len = unsigned_varint::encode::usize(buf.len(), &mut len_buf);
+    stream.write_all(encoded_len).await?;
+
+    // Write protobuf bytes
+    stream.write_all(&buf).await?;
+    stream.flush().await


I haven't tried it yet but I think following is equivalent:

Suggested change

// Encode message to protobuf bytes

let mut buf = Vec::with_capacity(msg.encoded_len());

msg.encode(&mut buf)

.map_err(|e| io::Error::new(io::ErrorKind::InvalidData, e))?;

// Write unsigned varint length prefix

let mut len_buf = unsigned_varint::encode::usize_buffer();

let encoded_len = unsigned_varint::encode::usize(buf.len(), &mut len_buf);

stream.write_all(encoded_len).await?;

// Write protobuf bytes

stream.write_all(&buf).await?;

stream.flush().await

let mut bb = bytes::BytesMut::with_capacity(MAX_MESSAGE_SIZE);

request.encode_length_delimited(&mut bb);

stream.write_all(&bb).await?;

Maybe we could add a test for how the encoded data should look like?

I think it's not right way to do, since stream could have mulltiple requests and when you read it to end you will skip them

This is for writing so it should not be problematic. As for reading I agree, the stream might contain multiple encoded messages. The prost library contains some helper functions to read the lengths; we might be able to drop the unsigned-varint altogether.

emlautarom1 · 2026-01-12T21:20:45Z

+    // Read unsigned varint length prefix
+    let msg_len = read_usize(&mut *stream).await.map_err(|e| match e {
+        unsigned_varint::io::ReadError::Io(io_err) => io_err,
+        other => io::Error::new(io::ErrorKind::InvalidData, other),
+    })?;
+
+    if msg_len > MAX_MESSAGE_SIZE {
+        return Err(io::Error::new(
+            io::ErrorKind::InvalidData,
+            format!("message too large: {msg_len} bytes (max: {MAX_MESSAGE_SIZE})"),
+        ));
+    }
+
+    // Read exactly `msg_len` protobuf bytes
+    let mut buf = vec![0u8; msg_len];
+    stream.read_exact(&mut buf).await?;
+
+    // Unmarshal protobuf
+    M::decode(&buf[..]).map_err(|e| io::Error::new(io::ErrorKind::InvalidData, e))


Same here:

Suggested change

// Read unsigned varint length prefix

let msg_len = read_usize(&mut *stream).await.map_err(|e| match e {

unsigned_varint::io::ReadError::Io(io_err) => io_err,

other => io::Error::new(io::ErrorKind::InvalidData, other),

})?;

if msg_len > MAX_MESSAGE_SIZE {

return Err(io::Error::new(

io::ErrorKind::InvalidData,

format!("message too large: {msg_len} bytes (max: {MAX_MESSAGE_SIZE})"),

));

}

// Read exactly `msg_len` protobuf bytes

let mut buf = vec![0u8; msg_len];

stream.read_exact(&mut buf).await?;

// Unmarshal protobuf

M::decode(&buf[..]).map_err(|e| io::Error::new(io::ErrorKind::InvalidData, e))

// Read unsigned varint length prefix

let mut buffer = Vec::with_capacity(MAX_MESSAGE_SIZE);

stream.read_to_end(&mut buffer).await?;

M::decode_length_delimited(buffer.as_slice())

emlautarom1 · 2026-01-12T21:24:09Z

+pub async fn send_peer_info(
+    mut stream: Stream,
+    request: &PeerInfo,
+) -> io::Result<(Stream, PeerInfo)> {


Could we use the following signature?

Suggested change

pub async fn send_peer_info(

mut stream: Stream,

request: &PeerInfo,

) -> io::Result<(Stream, PeerInfo)> {

pub async fn send_peer_info(stream: &mut Stream, request: &PeerInfo) -> io::Result<PeerInfo> {

No, since it will run as a future which should own a stream, otherwise it could lead to lifetime issues

emlautarom1 · 2026-01-15T20:08:37Z

+const DEFAULT_INTERVAL: Duration = Duration::from_secs(60);
+
+/// Default timeout for peer info requests.
+const DEFAULT_TIMEOUT: Duration = Duration::from_secs(20);


From where are we getting these constants?

For interval: https://github.com/ObolNetwork/charon/blob/54ec4d915806130b824e5f8497cf074cdfbcedfa/app/peerinfo/peerinfo.go#L29

Timeout was set by me

emlautarom1 · 2026-01-15T20:14:09Z

+    }
+}
+
+impl ConnectionHandler for Handler {


This is my biggest gripe with the approach: I have no way to review this code or assert that it's correct 😢

We can setup a meeting and I will make a walkthrough the project if you want, but it's the only way to go. Also, it's the easiest module for us to implement among all others, so the sooner we start using native libp2p approach - the better. This approach provides the best code maintainability and resilience

emlautarom1

After some internal discussion we'll move forward with this PR. A few things to keep in mind:

We need to eventually test edge case conditions like connections dropping mid message, peers disconecting and spamming peerinfo (potential unlimited vector growth)
Add references to the ping which was used as a base for this implementation'
Include some tests that validate peerinfo serialization/deserialization (ex. write multiple consecutive peerinfos to a stream and deserialize them)

varex83 added 25 commits December 5, 2025 15:41

feat: add tracing

913b56b

fix: linting, cargo-deny

51a691f

Merge remote-tracking branch 'origin/main' into bohdan/tracing

e3ece2e

feat: add p2p config

85a9b9c

fix: remove comment

65d5ae7

Merge remote-tracking branch 'origin/main' into bohdan/p2p-config

a312796

feat: add p2p metrics

1cc2a29

fix: rename label in example

c432d86

fix: linting

aeaa971

Merge branch 'bohdan/p2p-config' into bohdan/p2p-create-node

65d3a49

feat: [wip] add relay/ping

a82bbc7

feat: add behaviours module, add relay client support

a2fe791

fix: linter warnings

30a550b

Merge remote-tracking branch 'origin/main' into bohdan/p2p-create-node

df9fdce

fix: remove wildcard dependencies

ddbb57b

feat: add connection gater

3c5d3fe

fix: linter warnings

c663f61

fix: linter warnings

a244063

feat: create behaviour builder

adfc1ef

feat: add relay server functionality

31c4e27

chore: fix linter, polish code

152fbcf

fix: cargo deny dependency issue

529d794

chore: add docs

1eb752e

refactor: move relay-server to a different crate

7b72369

feat: add peerinfo (mvp)

ee147d0

varex83 changed the title ~~feat: add tracing~~ feat: add peerinfo Jan 5, 2026

fix: remove non-relevant comment

351db0e

Base automatically changed from bohdan/p2p-relay-server to main January 13, 2026 15:43

emlautarom1 reviewed Jan 15, 2026

View reviewed changes

emlautarom1 approved these changes Jan 19, 2026

View reviewed changes

varex83 added 4 commits January 20, 2026 12:00

Merge remote-tracking branch 'origin/main' into bohdan/p2p-peerinfo

5b13d50

fix: move dependencies to workspace

ef8dae3

feat: add compatibility tests

e6944a1

Merge remote-tracking branch 'origin/main' into bohdan/p2p-peerinfo

b60454a

varex83 merged commit 9ad5a5a into main Jan 21, 2026
7 checks passed

varex83 deleted the bohdan/p2p-peerinfo branch January 21, 2026 12:07

		futures = "0.3"
		futures-timer = "3.0"

Conversation

varex83 commented Jan 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions Bot commented Jan 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

emlautarom1 left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

emlautarom1 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

varex83 commented Jan 5, 2026 •

edited

Loading

github-actions Bot commented Jan 5, 2026 •

edited

Loading