Trajectory Planner using Ruckig Lib

If you still keep core dump, i can edit the appimage to run Using POSIX non realtime

How do I do that?

Yup, this is non RT kernel, will try on a RT kernel later.

Here is my proposal to use a spline in gcode.

You probably know how it is with GCode and Splines. Decades ago, when memory was expensive, they were interesting because they reduce file size (as do arcs). As discussed a few pages back on this thread, modern multiaxis CAM/ Postprocessors produce G1 line segments only, they don't even bother with G2/G3 arcs (R vs IJK, Absolute vs Relative). Since processing power and memory have become so cheap the interest in splines has all but disappeared. That is why nobody really uses them. If you asked somebody to create a postprocessor that uses splines, I guarantee the question would be: WHY?
So, unless you have found a new and compelling reason for using splines instead of lines (and arcs) you are flogging a dead horse.     

However we need a 4pnt spline to do the exotic path blending fillets where the clothoids are out of plane, or invalid.

That is what I suspected. I don't think there would be many objections to modifying spline syntax if it helped getting jolt limiting path blending for multi axis gcode working.

I'm afraid I can't help you with implementing the linked 3d clothoid algorithm.