Lately I've been trying to get to the bottom of a bug that has been happening but it seems like it should not be possible. Today I had about an hour of run time with the problematic hardware in which the bug was not observed despite no code changes since the last time I saw it. The only difference I can think of is that the hardware was plugged into a different USB port and I will be annoyed if that turns out to be a reproducible trigger for the behavior.