How do you handle a situation where a client insists a software bug exists, but your testing shows the software is performing as designed?
