Yes it is possible. Since types of the Constraint genus are finite sets of atom-type constraints, you can easily verify their equality.
However, a PolyKinds extension PolyKinds not required. In addition, there are very few situations where this kind equality would really be useful, because I do not see a practical way to pass polymorphic constraints as arguments c1 , c2 to Bla , so equality of constraints would be a tautology in each case ( Show ~ Show here) :
{-
dflemstr
source share