Compilation error "Could not deduce (FunctorConstraint [] a, "... about constraintkinds HOT 5 CLOSED

I'm on ghc 7.6 and it works fine. IIRC, the traverse class is not actually required for the functionality in hlearn-distributions. So if you can't upgrade, you could download the repository, remove that from the cabal file, install, then download hlearn-distributions like normal.

from constraintkinds.

I have a different issue now. I upgraded to ghc 7.6.1 and installed ConstraintKinds. Thanks! But now I'm having trouble with the Gaussian model in hlearn-distributions. Perhaps I should report this bug to the github page for hlearn-distributions. Anyway, it can't find Unbox'. This is version 0.0.1.1, as found via the cabal.

[3 of 4] Compiling HLearn.Models.Distributions.Gaussian ( src/HLearn/Models/Distributions/Gaussian.lhs, dist/build/HLearn/Models/Distributions/Gaussian.o )

src/HLearn/Models/Distributions/Gaussian.lhs:187:33:
    Not in scope: type constructor or class Unbox'
    Perhaps you meant `U.Unbox' (imported from Data.Vector.Unboxed)

from constraintkinds.

Interesting. Hackage is getting the same error when it compiles it, but on my system I can do

cabal install HLearn-distributions

and everything works just fine. There must be some dependency snag that I have installed, that didn't make it into the cabal file, but for some reason cabal was still willing to accept that all the dependencies were met.

I'll look into this in more detail, but this coming week is really busy for me so I can't promise when I'll figure it out.

from constraintkinds.

Someone else pointed out the problem was with an interface change in the vector-th-unbox package. I've updated the dependency list to reflect this, and re-uploaded to hackage. So hopefully is should work now.

from constraintkinds.

Excellent! That worked. Closing.

from constraintkinds.