Shared sequencer integration for fuel-core 0.40.0
#2451
Conversation
|
|
||
| let block = self.tendermint()?.block(height).await?; | ||
|
|
||
| for item in block.block.data.iter().skip(1) { |
There was a problem hiding this comment.
@Dentosal Added it, so I don't know=) I will remove this function for now, since it is not used
| pub tx_bytes: String, | ||
| } | ||
|
|
||
| pub async fn estimate_transaction( |
There was a problem hiding this comment.
do we want any retry logic here? can be a follow up PR
There was a problem hiding this comment.
The current implementation doesn't care if we are not able to submit the blob. We just try to do that, and if not, then it's okay.
In the future(follow-up issue), we need to worry about it and guarantee that all blobs are posted and there are no gaps.
|
|
||
| impl Client { | ||
| /// Create a new shared sequencer client from config. | ||
| pub async fn new(endpoints: Endpoints, topic: [u8; 32]) -> anyhow::Result<Self> { |
There was a problem hiding this comment.
i think it should also accept the expected client_id so that we don't misconfigure with the wrong network/chain.
on L68 we can perform the check
There was a problem hiding this comment.
What is the client_id? Do you mean chain_id? I want to keep the current implementation as simple, as possible.
Plus, if the RPC is malicious, it can lie about its chain_id
| .make_payload(timeout_height, fee, signer, account, order, topic, data) | ||
| .await?; | ||
|
|
||
| let r = self.tendermint()?.broadcast_tx_sync(payload).await?; |
There was a problem hiding this comment.
maybe we should have some retry logic here?
There was a problem hiding this comment.
The same logic here #2451 (comment)
|
|
||
| /// Sign a blob of data | ||
| pub async fn sign(&self, data: &[u8]) -> anyhow::Result<Signature> { | ||
| self.sign_message(Message::new(data)).await |
There was a problem hiding this comment.
perhaps we should add a check here for the length of the message?
we should return an error if the payload is of length 0, it's behaviour is undefined in the specs of the elliptic curves afaiu
There was a problem hiding this comment.
The Message performs hashing inside, so it will be just empty hash.
|
|
||
| let k256_public_key = | ||
| k256::PublicKey::from_public_key_der(cached_public_key_bytes)?; | ||
| Ok(Some(k256_public_key.into())) |
There was a problem hiding this comment.
q: why don't we just set the public key into the enum variant instead of its raw form? we won't have to recalculate it when verifying_key() is called every time.
as i understand the code, make_payload() is leading to this function being called repeatedly
There was a problem hiding this comment.
It is more question to @Dentosal =) Performance is not critical in make_payload because we call this function twice each 12 seconds
Description
This PR duplicates #1922, but on top of the
fuel-core 0.40.0.Since the current
masteris not ready to be released, but we want this feature in the testnet, we will create a minor release with added off-chain logic that posts data to the SS.All changes are the same as in the original PR, except
reqwestis bumped to0.12, because it has a fix important for SS.Before requesting review