plutovg VS cglm

Sub-pixel positioned anti-aliased lines and dots are simple. Extending that idea to drawing any shape is far more fiddly. I found this excellent rasteriser and this article which explains how it works. PlutoVG wraps that rasteriser in an easy-to-use API. I'm very happy with the results.

It's too bad you are using C. If you were using C++, you could use teh fantastic GLM library, which implements the vector math stuff from GLSL as CPU/C++ code. you might want to have a look at CLGM: https://github.com/recp/cglm

There's cglm, but I don't know how good it is. Really though for hobby OpenGL stuff, it's going to be 3x3, 4x4, 1x4 matrices, so rolling one's own stuff is not that terrible, if it proves to be a bottleneck (probably won't), you can drop in some fancy SIMD bullshit later.

There is a cglm: repo

Check out https://github.com/recp/cglm I use it judiciously.

GLM and CGLM should probably be mentioned as well since you'll probably want a good linear algebra library if you are going with just a rendering library instead of a full game engine.

cglm has two options for representing a vector: using a struct and using an array. The default one is using an array, but I'm not seeing what are the benefits of using an array instead of a struct. When using an array, you need to use out parameters, which make the code much more ugly when doing multiple operations. Using a struct, I can simply return the struct from the function. The only advantage of using an array would be that it's easier to iterate through it, but I could do the same using structs like this

It has been a while since I've written any C so it's very likely I'm missing something very simple. But I'm using CGLM and I have this simple program:

https://github.com/recp/cglm is a C version of GLM