use_sim_time Causes Very Long Sleep

[erichlf]

In the process of doing a bunch of things during the day I had set use_sim_time:=true and then later on I ran rosbaglive. When doing this it appeared that rosbaglive was not actually playing the data back. After some debugging it turns out that I was stuck at the line where rosbaglive sleeps until it is appropriate to play the next message. It turns out that this will be an extremely long time, since the first call to rospy.Time.now() will be 0 if use_sim_time is true. I think it would be good to check if this parameter is set and then at least log a warning, so that the user know they will likely wait for the rest of their life before the next message is played.

[DLu]

This package hasn't had development in a good long while. I'm glad you've found use for it.

Please feel free to make a PR with the proper warning message.