Couldn't match kind '* -> *' with * about beam

[tathougies]

This seems like a bug in GHC. Does it work in 8.2?

[3noch]

Unfortunately I won't be able to test that for a while. I wouldn't be surprised if you're right.

[tathougies]

Do you have any example I could try to reproduce?

[vwwv]

I've got a similar issue a couple of times, where ghc was complaining Couldn't match kind '* -> *' with '*' then showing exactly the same types as expected and actual. In my case, It ended up being something totally unrelated to beam:

In one case, by error, I used a constraint (* -> Constraint) in a type application expecting a type (*), in the other case I was using a map (* -> * -> *) in a list comprehension instead of a list (* -> *).

In both cases, the error was marked on a different line it actually happened, and over a type that has not much to do with it, making it something really frustrating to debug even though they were pretty dumb bugs. I think the problem, somehow, on some ghc versions, ghc's error report heuristic doesn't play well in some situations where beam's columnar approach (i.e C f Foo) is involved.

In my case, I was using ghc-8.0.2, maybe this has improved in later ghc versions?