chaindexing / chaindexing-rs Goto Github PK

The engine should be able to intelligently prune processed and stale events.

Currently, we make blocking RPC calls to fetch blocks for ingested events. This could be a performance bottleneck for DApps that do not require block_timestamp values. It would be better to have an opt-out configuration to allow drastic latency reduction in such DApps.

Currently, this library uses tokio-postgres and diesel as the RDBMS clients. We need to pick one for obvious reasons.

ANALYSIS (WIP)

This issue lists Renovate updates and detected dependencies. Read the Dependency Dashboard docs to learn more.

This repository currently has no open or pending branches.

Detected dependencies

• Check this box to trigger a request for Renovate to run again on this repository

Action: Investigate tokio-postgres.

Analysis: Timestamp fields should be encoded as serde::Value::String("\"<value>\"") -- the inner extra quotes are essential for serde to reach an agreement with chrono when deserializing. However, it seems tokio-postgres somehow gets rid of those extra quotes when converting a Row to its equivalent serde::Value.

Currently, blocks are fetched for the logs in a blocking fashion. It will greatly improve the ingestion performance if we fetched them concurrently instead.

This will help guide users when they have written an invalid SQL syntax. To stay language agnostic, it must be runtime-based. Eventually, after most libraries have been built, we can expose an extra macro, or even sqlx's parser, to be used at compile time.