Azimuth window not taken into account when calculating point cloud timestamp in ROS_TIME mode

[AlessandroBiaggi]

Describe the bug
The ouster_ros::LidarPacketHandler::lidar_handler_ros_time method incorrectly assumes that the azimuth window always includes 0°. This leads to an erroneous lidar_handler_ros_time_frame_ts timestamp when the azimuth window does not pass through 0°, as it incorrectly assumes that the last columns of the previous scan and the first of the next are in the same packet.

Additional Information:

• git commit hash: 8833bee

[Samahu]

yeah I agree this is a bug, I am actually aware of it but I it has been low on my list of TODOs. But you are absolutely right the code assumes you are using the full 360° FOV.