A while back I shared my first-ever game in Rust—a Dangerous Dave clone built using Macroquad and Tiled as a way to learn the language and survive the compiler's ownership rules.

https://www.reddit.com/r/rust_gamedev/comments/1jpt3e2/dangerous_dave_rust/

I wanted to tackle something a bit more ambitious for my second project: a 2D dungeon roguelike I've been calling Catacombs. Since I wanted to understand how everything works under the hood rather than relying on an library or game engine, I've put together a custom stack using raw OpenGL and a handful of modular crates.

The Tech Stack

• Graphics & Windowing: glow coupled with glutin and winit for context creation and window management, using nalgebra-glm for matrix math and projections.
• ECS & Scheduling: hecs paired with hecs-schedule for managing components and system execution order.
• UI & Layout: egui (integrated via egui_glow) for tooling and debug overlays, alongside taffy for flexbox-based UI layout calculations.
• Levels & Audio: ldtk_rust for loading LDtk level designs and libfmod for handling audio middleware.
• Data & Concurrency: serde and ron for data-driven asset and configuration serialization, with rayon, crossbeam, and parking_lot handling concurrency and synchronization.

Graphics

Bought the Dungeons & Pixels pack from

https://indie-vova.itch.io/dungeons-and-pixels-starter-pack

AI

About %85 of the code was written from scratch, learned a lot, then I've had access to an AI tool that did code reviews for me, fixed the code manually, few of the suggestions didn't make much sense to me so I have ignored them.

I have generated the start up page using Google's Gemini free credit.

What's Next?

Since I never really built out a story or a concrete endgame vision for Catacombs, I've decided to officially wrap up the project here and archive the repo. I have also learning a lot and improved my Rust programming skills, I can see that transition from the Dangerous Dave game code to Catacombs code.

For my next steps on this Rust learning journey, I'll be moving past OpenGL to build a low-level renderer using Ash for Vulkan, with plans to experiment further with advanced procedural generation algorithms and explicit graphics pipeline design from scratch.