Nicolas Pouillard

I'm pretty confident we won't need any migration \o/

Since ClassName is a newtype the data representation didn't change it should be possible to make it work without any migration. I would test this by writing an instance of...

I will experiment with the safecopy instance first then.

What impact would this have to the interfaces? I used to be in favor of a change to help pass the current user around but not anymore. For the `currentUserAddDb`...

I was indeed expecting new wrappers such as the two proposed by @andorp. However one might need at least one more to combine two updates into one. I think that...

- CSRF protection has been merged in #490 - Cookie encryption could be done with https://hackage.haskell.org/package/wai-session-clientsession (or equivalent)

Random (pseudo-random) `UserId` is actually a good idea. In particular it hides how many users are subscribed and when did a particular user approximately subscribed.

:+1:

Helping to get a deterministic build of docker images would be great.

For the record here is advanced API used in Haskell: [Data.Map.Merge.Strict](https://hackage.haskell.org/package/containers-0.6.2.1/docs/Data-Map-Merge-Strict.html). Would such an API be considered if implemented?