Code Review Series - Unreal Engine Developer

Introduction

• Presenter: TH
• Context: Code review series where viewers send code for review.
• Current review of a 23-year-old developer's code, using a virtual machine (VM) for safety.
• Developer's background: Works with Unreal Engine, wanted to improve C++ skills by building a game engine from scratch.

Game Engine Overview

• Developer: Axel
• Libraries used: Glad and STB image.
• Platform-specific: Only supports Windows, no Mac or Linux support.
• The game engine project is named "Greg engine."
• Example game: "Basket Man" - a simple game included in the project.

Educational Approach

• Building a game engine from scratch is educational but resource-intensive.
• Not using third-party libraries can limit progress but provides deep learning.

Code Structure and Architecture

• Project Organization: The project consists of a static library (Greg engine) and an executable (Basket Man).
• Level Management:
  • Levels are stored as unique pointers in the engine and returned as references.
  • Concerns about using raw pointers and references instead of shared pointers for better resource management.
• Component System:
  • Components are managed as unique pointers, which could lead to fragmented memory.
  • Suggestion to store components as continuous blocks of memory for better performance.

Window and Rendering

• Window Management:
  • Uses low-level Windows API for window management.
  • Custom windowing code is typical for educational purposes but not for production.
• Rendering:
  • Current rendering uses individual vertex arrays for each sprite, which is not efficient.
  • Recommendation to use a single vertex array setup with transformation matrices for efficiency.

Math Library

• Developer provided custom matrix classes and math functionality.
• Suggestion to use established libraries like GLM for better performance and accuracy.

Code Review Observations

• Some static functions and data structures that should ideally be instance-based.
• Use of static physics and camera systems, which might be better as level-specific instances.
• Issues with warnings and errors due to missing includes and data type mismatches.
• Some unconventional approaches to memory management and data structuring.

Rendering Performance Concerns

• Inefficiencies noted in sprite rendering due to repeated allocations and setups.
• Importance of batch rendering techniques for performance improvement.

Conclusion

• Overall, the code serves as a strong educational tool.
• Suggestions for improvement in terms of architecture, use of libraries, and rendering strategy.
• Encouragement to continue exploring and refining the engine for learning purposes.

Additional Resources

• Mention of Code Rabbit AI, a tool for automated code review, suitable for learning and project development.