How did you run the config generation? This is what I get:

#c.Authenticator.whitelist = set()

which is correct.

Hi @hokiegeek2,

Thanks for the issue report. I'm getting the same results as @minrk for Python 3.5 (as a reminder JupyterHub runs on Python3). Out of curiousity, how did you install JupyterHub? Perhaps there was already a config file in your environment.

I'm going to go ahead and close this issue since we can't replicate it. If you have additional information, I would be happy to look into it.

Thanks! - Carol

Hey Everyone,

I apologize for the delay in following up on this. @minrk Yes, the config generation results in this format. However, when I manually set the users in this data structure:
c.Authenticator.whitelist = set('hokiegeek2')
I get this message:
[W 2016-12-03 23:25:36.672 JupyterHub auth:138] User 'hokiegeek2' not in whitelist.

When I switch to this, the whitelist filter works correctly:
c.Authenticator.whitelist = {'hokiegeek2'}

--John

Thanks for the update @hokiegeek2.

@minrk Do we want to edit the comment about whitelist and/or add a note in the docs?

set('hokiegeek2'), like list('hokiegeek2'), treats the string as a collection and transforms each item (character) in the string to an item in the set, so what you get is {'h', 'o', 'k', 'i', ...} instead of {'hokiegeek1'}. If you wanted to create that single-item set with the set() constructor, you would do set(['hokiegeek1']).

I've opened #885 to give a more informative message when it sees that a whitelist contains single-character usernames.

Thanks @minrk 🍪