Better error classification in clients/feeder

[ashu26jha]

Problem

Currently, if we try to fetch the block (lets say 1,000,000) from the feeder we get 400 Bad Request which is a generic error. Meanwhile the feeder gateway shows:

{"code": "StarknetErrorCode.BLOCK_NOT_FOUND", "message": "Block number 1000000 was not found."}

We want the error message to reflect the response.

Advantages

1. This will help in debugging.
2. This will be needed to improve reorg handling

[ashu26jha]

#1820

[anubhavitis]

I am applying to this issue via OnlyDust platform.

My background and how it can be leveraged

I have over 2 years of experience in building Go-based applications, having worked with companies like CoinSwitch and currently Fileverse. I have designed, built, and deployed services from scratch that cater to hundreds of thousands of users daily. My expertise spans supporting both Web2 applications and Web3 dApps, making me well-equipped to contribute effectively to your project

How I plan on tackling this issue

Things that are on top of my mind:

• Review the code where the block fetch request is made and where the error is being handled.
• Locate the part of the code that translates errors from the feeder gateway into HTTP responses.
• Update the error handling logic to capture specific error responses from the feeder gateway.
• Modify the HTTP response to include the specific error message from the feeder gateway.
• tests: unit tests and over staging/dev environments