Change bumper collisions to account for rubber

[sleddog]

Currently the bumper collisions are treated as elastic collisions just like striking a mirror.

        ball = this.balls[i];
        //if potted, no collision with bumper
        if(!ball.potted) {
            if(ball.position.x > X_MAX-PREFS.bumperWidth-ball.radius || ball.position.x < 0+PREFS.bumperWidth+ball.radius) {
                    ball.velocity.x *= -1;
            }
            if(ball.position.y > Y_MAX-PREFS.bumperWidth-ball.radius || ball.position.y < 0+PREFS.bumperWidth+ball.radius) {
                    ball.velocity.y *= -1;
            }
        }
        ball.update();

Note the velocity is simply flipped. We should account for some velocity decrease after striking the rubber bumper. Also make this rubber bounce be a preference.

[kinnerapriyap]

Hey! I'd like to take this up, if that's okay.
What coeff of restitution should I use with the rubber bumper and ball?
I'm not sure what you mean about making the rubber bounce a preference, so if you could explain that out a little bit, it'd be great.

[sleddog]

If you press H the preference dialog appears. Note other settings like ball mass, friction, etc... the coefficient of restitution could be one of these settings. Not sure what value to use but some experimentation would be needed. Thanks in advance if you take this ticket @kinnerapriyap !

[kinnerapriyap]

I put in a range up to 5 for bumperBounce (least to most bounce) because you've used such a range for the rest of the preferences. Do let me know if you'd like a different range.
Using KE loss as a preference was a complication, because it would need to account to negative cases also. This seemed like a better option in terms of bounce visualisation while selecting preferences.

[sleddog]

@kinnerapriyap brining over your comment from the closed PR:
"The angle if off because the velocity parallel to the bumper has no force on it and remains the same. The velocity perpendicular is reduced because of loss in kinetic energy due to collision. So since only one component of velocity changes, the angle changes. Do let me know if you want me to make any changes to it."

I think if the angle stayed the same (or wasn't as a pronounced change) and only the velocity changed would be best. If you want to continue this, can you reopen the PR and tweak this behavior?

Playing in real life the ball collisions with bumpers does cause an angle change based on speed but it is very minor.