That version of pylink does not include pylink.EyelinkCustomDisplay, which is used by the code I have to talk to the Eyelink 1000. I think @jon at one point said that pylink 0.3.1 had mistakenly been included in some versions of Psychopy, so maybe that is the issue.
Looking at the SR Research Eyelink forum, it seems my eyetracker communication code is out of date as the examples they currently provide don’t use pylink.EyelinkCustomDisplay. So perhaps my problem won’t affect others, and I should try updating my code. Not sure when I’ll get to that, but don’t you worry about it; the ball’s now in my court.