Window::get_size() should return (u32, u32)

[atomicbitty]

The glfw.create_window() function uses u32 for the width and height, so I think it would make more sense to return a pair of u32s in Window::get_size() instead of a pair of i32s

[aloucks]

glfwGetWindowSize returns values as i32 (as well as all of the other *size functions and events). Someone would need to confirm that they can never hold negative values and then change all of them to u32 for consistency. Note that I'm fairly certain that WindowEvent::Pos(i32, i32) can have valid negative values and explicitly not on the list below.

• Window::get_frame_size
• Window::get_framebuffer_size
• Window::get_size
• Window::set_size
• Monitor::get_physical_size
• WindowEvent::Size
• WindowEvent::FramebufferSize

[RpxdYTX]

But Glfw::create_window accepts u32 for width and height instead of i32, which is what it is in the C version.

[aloucks]

But Glfw::create_window accepts u32 for width and height instead of i32, which is what it is in the C version.

The window can be resized.

All of the glfw functions (other than create) return or use signed int. Either this was an oversight in GLFW or there is some edge cases on some platform that could return negative numbers.

I think the best option here is to open an issue with GLFW. If they decide to change these to be unsigned values than we could update to match accordingly. If they don't want to change these to be unsigned, then we'll have our reason why.

In either case, we should continue to use the same type (currently signed int: i32) that glfw is using.