Finomnis

Yah, I figured. I'm just in the process of porting u8g2's font subsystem to Rust and stumbled across this, so I thought I'd report it here. But yah, I understand...

Please provide the exact input for both algorithms. They do not have a `width` and a `height` input, and the `(0,0)` and `(1,1)` values should clearly be identical when `aspect`...

My suspicion is confirmed by this minimal example: ```rust use std::f32::consts::PI; use nalgebra_glm::{perspective_rh_no, Mat4}; fn perspective(fov: f32, aspect: f32, z_near: f32, z_far: f32) -> Mat4 { debug_assert!(aspect > f32::MIN_POSITIVE, "");...

As a sidenote, they seem to be aware of this issue. They added an explicit warning to the documentation: https://docs.rs/nalgebra-glm/latest/nalgebra_glm/fn.perspective_rh_no.html#important-note

It seems this is not the first time this was brought up: https://github.com/dimforge/nalgebra/issues/423 Might be worth reopening or creating a new issue to remind them of fixing it at the...

@Makogan I deleted my comments as they were not relevant for the improvement of this crate.

I don't think there is much use to `Size` until you can actually do math with it. Right now you can't even create a `[0u8; W*H]` with that information. I...

What's your usecase? A future shouldn't become ready every now and then, it should become ready exactly once; reason is that it has to produce a value at some point,...

It even is a direct violation of the [poll()](https://doc.rust-lang.org/std/future/trait.Future.html#tymethod.poll) function: > Once a future has finished, clients should not poll it again.

> The question was: Why is it a TODO? If this is so clearly defined, then this should be a panic instead. I agree.