Hlib Kanunnikov
Hlib Kanunnikov
Also remove the TODO next App node instanitiation
## Context This EPIC is an entry point for the BlockSync Storage Overhaul efforts. For deeper context, look at ADR document: #1037 ## Tasks ### Preliminary Work - [x] #878...
The `fraud` pkg is designed as a generic library that encompasses p2p logic for any Fraud Proof. It allows any pkg depending on it to define the custom implementation of...
The latest rsmt2d version allows us to set cells/share directly into the imported square. Hence, we can avoid using flattened slices and the complexity it introduces when calculating position or...
The problem comes after recent node refactoring and is located in [using start context](https://github.com/celestiaorg/celestia-node/blob/main/nodebuilder/header/module.go#L61-L62) as lifecycle context. The start context is short-lived and is canceled after the node is started....
Introduced #1104 and #1105 are allocations and CPU-intensive serialization functions. We should know how they perform and always have a starting point on where to improve further.
Fraud proofs are expected to be big. We need to ensure that gossiping and syncing are reliable for 2mb+ proof sizes
Instead, each protocol like HeaderEx should have an option or config to receive and prepend a chain ID in the constructor and not during global initialization.