@jakub_neruda Same, I'm quite liking the flexibility that it gives, even if it is harder and more time consuming making it from the ground up

I'm using C++ and DirectX11, with minimal libraries added, so far I'm only using imgui (for debug UIs) and DirectXTK (but so far only used the image loading functions from it). Mostly not using many libraries to learn about how things can be made and have everything more tightly integrated together. and one of the reasons I'm making an engine was to learn

@Elusive_Fluffy Good approach, I apply the same mindset to my hobby gamedev projects as well 👍

I am using SFML+TGUI for my projects since it provides good primitives for audio, video, ui and network, and I am building my frameworks atop of that.

Except SFML is a 2D library (unless you want to interface with OpenGL directly) and I am planning to build a 3D renderer using 2D facilities once I am done with my current project 😅

@jakub_neruda Oh, nice, that sounds like a good foundation. I hadn't really heard of SFML or TGUI before

That sounds like a tough challenge to build a 3D renderer like that, good luck with it when you get around to it

Also just realised another library I'm using is xinput for controller input, I'll probably use a library for audio too, maybe will use something like openAL

@Elusive_Fluffy SDL2 and SFML are the go-to C++ 2D libraries that both should allow you access their OpenGL context to do whatever rendering shenanigans you want (in case their rendering API is not enough for you).

They already integrate XInput and OpenAL (and others) to give you multiplatform (and convenient) API for those primitives that might work differently on each platform. I am not even sure if XInput can deal with all Windows-compatible controllers (Dual Sense certainly comes to mind) - that is SFMLs job to handle.

What I love about SFML is also the fact that I can very easily cross-compile between Windows and Android (see https://nerudaj.itch.io/magrider and https://github.com/nerudaj/MagRider if interested).

As with regards to 3D renderer, I am pretty confident it is possible. I've already solved perspective-correct texture mapping with shaders before and otherwise it should be "just" a triangle conversion pipeline from world space to screen space. I am mainly worried about the performace, so fingers crossed 😉