[Tech]: Improve fixed numbers calculation precision

[vovac12]

Problem statement

Prevent overflow in fixed number calculations

Description

The current implementation uses i128 as the base for fixed number calculations. In certain cases, this leads to overflow during intermediate calculations, causing inaccuracies or failures in the computation.

To resolve this issue, we should replace i128 with 256-bit numbers (e.g., U256 or similar) for intermediate calculations. After the calculations are complete, results can be safely converted back to u128 for storage as balances

Definition of Done

• Create fixed number type with 256 bit number as base
• Replace fixed number type in critical parts of XYK pool calculations

Requirements

• External interfaces should remain the same
• Try to avoid of using types with dynamic allocation