Use modern OpenGL

[v1993]

Is your feature request related to a problem? Please describe.
If we want to make it possible to port cubosphere to mobile world, using OpenGL 3.1+ is simply a requiment.

This will be huge work, and lua shaders will be just gone, but it is just required.

Describe the solution you'd like
Replace Fixed Pipeline with Programmable one, i.e. GL shaders + VBO.

Describe alternatives you've considered
There are two of them:

1. Throw out cubosphere and write cubosphere 2.0 from int main. Looking at all arhitecture flaws and problems, it is probably a right way…
2. Just ignore entire problem and use OpenGL 2.0.

Additional context
Nope.

[qwertychouskie]

Actually, for best compatibility for older Android devices, GLES 2 support is important.

[ghost]

Personally, I'd go with rewriting the whole thing. It would be more rewarding than putting it off until OpenGL 2 is basically dead (or later).

The old PC in my bedroom won't like it either way.

[v1993]

@cddepppp256 Yeah, I completely agree with you. To achive better compatiblity with mobile devices (which have very limited shader support) and overall portability across modern ones I seriously think about using vulkan. It also will be very useful for me personally to learn it.

[qwertychouskie]

The number of Android devices that support Vulkan is quite small (https://www.quora.com/What-kind-of-Android-devices-are-compatible-with-Vulkan). Honestly for supporting the most devices OpenGL ES 2.0 is necessary at least as a fall-back renderer. OpenGL ES 3.0 will also run on a decent chunk of devices, but just because a device supports the API doesn't mean it is powerful enough to run a full-fat rendering engine.

In short, the quickest way to get Cubosphere running on the most devices is to tweak the current renderer to also support OpenGL ES 2.0.

P.S. I just had a crazy idea: Cubosphere using WebAssembly and WebGL, running in a browser :)

[v1993]

@qwertychouskie Tweaking for OpenGL 3.0 is very, very hard to do. You can try, but overall work is so tedious that I'm not going to do it at all. I don't think that modifying such overall bad designed project that much is good idea, honestly.

As for Vulkan, there are libs which allow using it with OpenGL as backend.