>this month
Working on Object Oriented Bounding Box collision. But the [i]simple high-school level math[/i] was too hard for my elite hacker brain, so I am doing planes ("half-spaces") instead. This roughly doubles the amount of math for the CPU but also the code is less confusing. (Actually I think the problem's been floating point precision.) My intent is basically to hard-code reasonable behavior. Yesterday I got sliding-along-the-plane-when-its-at-an-angle working. I've had elevator-type motion working, and thought to try pushy-pushy motion. You stick to it which is reasonable for an elevator but not for pulling. But a more serious problem is phasing through it as depicted here. But the endless sinusoidal motion, it gave me nausea. So today I work in debug cam.


>>105753097
I've been sorely tempted to replace nalgebra. I'll probably do it eventually.
>whywouldyouwanttoreplacenalgebra.WEBP
(gdb) print t
$1 = nalgebra::base::matrix::Matrix<f32, nalgebra::base::dimension::Const<3>, nalgebra::base::dimension::Const<1>, nalgebra::base::array_storage::ArrayStorage<f32, 3, 1>> {data: nalgebra::base::array_storage::ArrayStorage<f32, 3, 1> ([[1.97101212, 2.67225742,
-2.5]]), _phantoms: core::marker::PhantomData<(f32, nalgebra::base::dimension::Const<3>, nalgebra::base::dimension::Const<1>)>}