aoc-runner-derive VS rust-sdl2

I don't think that a way do this is actually guaranteed to work, see Crate local state for procedural macros? Issue #44034. That said there are crates that do this anyway since currently rustc expands macros from top to bottom in source files and doesn't cache the result so the proc macros are always re-run each time the code is compiled. Two crates that rely on this by using global variables in their procedural macro code are: enum_dispatch and aoc-runner (see here).

I don't think that SDL and WGPU are really comparable. You'd be better off comparing SDL to something like Notan or even the SDL bindings for Rust. That said, SDL will have vastly more learning resources than Wgpu or Notan or any Rust alternative. If you know and like rust you can probably use the resources with the Rust SDL bindings.

Those are 3 very different libraries. Are you looking for something in the middle, or do you want all of the features you described? For what it's worth, there's SLD2 bindings for Rust https://github.com/Rust-SDL2/rust-sdl2

There's already a Rust binding for SDL2. See https://crates.io/crates/sdl2

nope. https://github.com/Rust-SDL2/rust-sdl2/issues/884 emscripten target won't work yet either.

it seems like what you are looking for is sdl2 bindings

Dude. It's literally at most 50 lines of code, the entire project. What kind of documentation do you expect me to write? The library I'm using is this one. I don't know why you'd need to know anything about it since it's purely something that can appear literally anywhere, it's part of the language.

*and* SDL2-sys *and* Rust-SDL2

The SDL2 crate had something like this, and it's an option I'm considering. However, if possible, I'd prefer a different, simpler GUI library, so if there's an input-specialized crate out there that does this, I'd love it.