the copy/paste that is Labeled and Data from Weighted has made it pretty clear that there's a need for either better abstraction and reuse between the graphs implementations, or some approach to code generation that allows these permutations to be written out automatically.
certainly the test system is amenable to such code generation, as it's designed to test behavioral correctness as a product of the implemented graph interfaces.
the copy/paste that is Labeled and Data from Weighted has made it pretty clear that there's a need for either better abstraction and reuse between the graphs implementations, or some approach to code generation that allows these permutations to be written out automatically.
certainly the test system is amenable to such code generation, as it's designed to test behavioral correctness as a product of the implemented graph interfaces.