v1: Initial implementation #1
Conversation
- Abstract db handling logic - Add mongodb implementation
- Keep track of last processed slot. - Refactor slot processing logic. - Handle `unwrap()` calls properly.
ci: fix CD workflow
src/beacon_chain/mod.rs
Outdated
| pub async fn get_block( | ||
| &self, | ||
| slot: Option<u32>, | ||
| ) -> Result<Option<Block>, Box<dyn error::Error>> { |
There was a problem hiding this comment.
If we wanna have fancier error handling (totally doable later down the line) we could probably use anyhow to easily bubble up/downcast errors and define all the possible error types for this API client with thiserror (here's a good example)
There was a problem hiding this comment.
I was having a bit of trouble finding the best approach for error handling.
This definitely seems the way to go. Thank you for sharing this crate and example!.
src/slots/mod.rs
Outdated
| utils::web3::{calculate_versioned_hash, get_eip_4844_tx, get_tx_versioned_hashes}, | ||
| }; | ||
|
|
||
| type StdErr = Box<dyn error::Error>; |
There was a problem hiding this comment.
again, recommend using anyhow :D
src/slots/mod.rs
Outdated
| let beacon_block = match beacon_api.get_block(Some(slot)).await? { | ||
| Some(block) => block, | ||
| None => { | ||
| info!( |
There was a problem hiding this comment.
nice! Have you looked into tracing? it's essentially logging on steroids—will help with instrumenting robust logging and we could eventually trace all API calls which will help debug stuff in the future.
There was a problem hiding this comment.
I found out about it yesterday when I was working on parallelizing the indexer.
The concept of structured spans, as opposed to conventional logs, makes it a more suitable choice for asynchronous and multi-threading environments.
Thank you for sharing!
- Move env variables logic to its own file. - Generalize std error.
- Use `reqwest` client. - Add try/catch instantiation - Rewrite functions to make it more idiomatic
Resolves #2 , resolves #8
Notes: