Rust port of https://github.com/JoeyDeVries/LearnOpenGL
You should be able to follow the tutorials at https://learnopengl.com/ with this - the code structure has been kept similar to the original C++ wherever possible.
This also means it's not necessarily the most idiomatic Rust code. For example, some standard naming convention lints are disabled and all OpenGL calls are "raw" and wrapped in
Run individual tutorials like this:
cargo run 1_3_2 (for
For reduced compilation times, you may only compile the code for a certain chapter by adding
--no-default-features --features chapter-1 for example.
- You can mostly ignore the setup instructions at Getting-started/Creating-a-window. Just create a new project with
cargoand copy the dependency section from Cargo.toml. Only
glfw-rsmight need some more setup, see here for details. You can also use glutin (a pure Rust alternative to GLFW), but the API is a bit different, so following the tutorials might not be as straight-forward.
- You might be tempted to use glium instead of raw OpenGL. I'd recommend against that, at least in the beginning, to get a good understanding of how OpenGL really works. Also, glium is not actively maintained at the moment.
- If you experience black screens or weird rendering artifacts, check out the
glCheckError!macro from chapter 7.
- exercises have been mostly omitted. You can look up the solutions in the original C++ source.
- For simplicity
tobjis used instead of
assimp(simpler interface, pure Rust and later tutorials only load OBJ files anyway). For alternatives see here and here.
imagecrate is quite slow in debug mode - loading the nanosuit textures takes so much time that it can be faster to use release mode (including compile time).
Status: partially done (4/9).
Status: partially done (1/2).
Debugging complete (the other two are not in the repo)
A note about the code organization
Originally each tutorial was a separate executable (using
cargo run --bin <name>. This didn't play very well with the
clippy (-> rust-lang-nursery/rls#132). Now all are integrated into the main binary, which leads to long compile times. As a workaround there are now feature flags for each chapter.