feat(net): add discv4 crate #113
Conversation
mattsse
added
C-enhancement
|
currently blocked by pending ethers-corre bump with breaking changes. |
|
Update: reimplemented the entire protocol TODO
|
Codecov Report
@@ Coverage Diff @@
## main #113 +/- ##
==========================================
- Coverage 72.50% 68.78% -3.73%
==========================================
Files 150 146 -4
Lines 10642 10575 -67
==========================================
- Hits 7716 7274 -442
- Misses 2926 3301 +375
📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more |
|
the self lookup step is a bit hacky atm, there's a bug in |
|
|
||
| #[tokio::test(flavor = "multi_thread")] | ||
| #[traced_test] | ||
| #[ignore] |
There was a problem hiding this comment.
is this so we don't get banned? i'd maybe feature gate it
There was a problem hiding this comment.
this is ignored since this connects with live nodes, started writing a mock service that allows us to test messages.
There was a problem hiding this comment.
sweet - let's also keep in mind a integration test runner which spawns geth nodes and peers with them (let's just copy-paste geth.rs from ethers-core and add options for p2p config)
There was a problem hiding this comment.
will take this on and track
| #[test] | ||
| fn test_noderecord_codec_ipv4() { | ||
| let mut rng = thread_rng(); | ||
| for _ in 0..100 { |
There was a problem hiding this comment.
wshould be replaced with proptest or test-fuzz, but let's do separately cc @joshieDo , lets track
| let signature: RecoverableSignature = SECP256K1.sign_ecdsa_recoverable( | ||
| &secp256k1::Message::from_slice(keccak256(&payload).as_ref()) | ||
| .expect("is correct MESSAGE_SIZE; qed"), | ||
| secret_key, | ||
| ); |
There was a problem hiding this comment.
lets track in an issue about maybe using k256 down the line if performance improves
There was a problem hiding this comment.
wanted to do this, but the other p2p stuff also relies on secp256k currently, so we should migrate this in one step. will track
| #[cfg(test)] | ||
| mod tests { | ||
| use super::*; | ||
| use crate::SAFE_MAX_DATAGRAM_NEIGHBOUR_RECORDS; | ||
| use bytes::BytesMut; | ||
| use rand::{thread_rng, Rng, RngCore}; | ||
|
|
||
| fn rng_endpoint(rng: &mut impl Rng) -> NodeEndpoint { |
There was a problem hiding this comment.
should also be done w fuzzer @joshieDo
There was a problem hiding this comment.
In essence this is an Arbitrary impl for NodeEndpoint since all values are randomly generated.
There was a problem hiding this comment.
test-fuzz does not require Arbitrary, only requires:
-
the fuzzed (input) type to be
SerializeandDeserialize -
a
testwhich can be run withcargo test
https://github.com/foundry-rs/reth/blob/main/crates/interfaces/src/db/codecs/fuzz/mod.rs#L47-L50 -
the above
testshould go through a function annotated with#[test_fuzz::test_fuzz]. Ideally the function should receive the targeted object.
https://github.com/foundry-rs/reth/blob/main/crates/interfaces/src/db/codecs/fuzz/mod.rs#L41-L45
| /// # use std::io; | ||
| /// use std::net::SocketAddr; | ||
| /// use std::str::FromStr; | ||
| /// use rand::thread_rng; | ||
| /// use secp256k1::SECP256K1; | ||
| /// use reth_discv4::{Discv4, Discv4Config, NodeId, NodeRecord}; | ||
| /// # async fn t() -> io::Result<()> { | ||
| /// // generate a (random) keypair | ||
| /// let mut rng = thread_rng(); | ||
| /// let (secret_key, pk) = SECP256K1.generate_keypair(&mut rng); | ||
| /// let id = NodeId::from_slice(&pk.serialize_uncompressed()[1..]); | ||
| /// | ||
| /// let socket = SocketAddr::from_str("0.0.0.0:0").unwrap(); | ||
| /// let local_enr = NodeRecord { | ||
| /// address: socket.ip(), | ||
| /// tcp_port: socket.port(), | ||
| /// udp_port: socket.port(), | ||
| /// id, | ||
| /// }; | ||
| /// let config = Discv4Config::default(); |
Add discv4 support.
This merges parts of https://github.com/vorot93/discv4 and parts of https://github.com/sigp/discv5 (kademlia) and also applies reth rlp encoding.
TODO