Possible bug and fix in GetGazeRay()

[JCorvinus]

I think I've found a bug in GetGazeRay(), specifically in the origin output. Look at the code below:

    if (gazeIndex == GazeIndex.COMBINE)
    {
        valid = eyesData[(int)GazeIndex.COMBINE].GetValidity(SingleEyeDataValidity.SINGLE_EYE_DATA_GAZE_DIRECTION_VALIDITY);
        if (valid)
        {
            origin = eyesData[(int)GazeIndex.COMBINE].gaze_origin_mm * 0.001f;
            direction = eyesData[(int)GazeIndex.COMBINE].gaze_direction_normalized;
            direction.x *= -1;
        }
    }
    else if (gazeIndex == GazeIndex.LEFT || gazeIndex == GazeIndex.RIGHT)
    {
        valid = eyesData[(int)gazeIndex].GetValidity(SingleEyeDataValidity.SINGLE_EYE_DATA_GAZE_DIRECTION_VALIDITY);
        if (valid)
        {
            origin = eyesData[(int)gazeIndex].gaze_origin_mm * 0.001f;
            direction = eyesData[(int)gazeIndex].gaze_direction_normalized;
            direction.x *= -1;
        }
    }
    
Notice that the direction vector's x component gets flipped as a coordinate system conversion, but the origin of the ray does not. This video shows that the ray's orientation gets applied to the opposite eye:

Video
For reference, Left is red, right is blue. You can see that the origin is coming from the closed eye, but the direction is coming from the open one

Applying the flip to the origin as well as the direction results in the following:

Video

@Daniel_Y @Andy.YC_Wang @Corvus @Jad

[HackPerception]

@JCorvinus - Thank you for your detailed report - we appreciate that you took the time to not only report this but to provide an video example of the bug and a suggested fix. I've tagged the relevant team members onto your post so they can look into the report.

Cheers!