Andrey Mokhov

> I have opened an issue here: haskell/haddock#827 to see what can be done. Awesome, thanks! > Which one do you prefer ? I prefer to stick to properties throughout...

This looks like an awesome library we should give a try: https://hackage.haskell.org/package/hedgehog-classes-0.1.1.0/docs/Hedgehog-Classes.html

This is indeed annoying, but I don't know how to nicely capture acyclicity in types. We could provide a separate function for `topSort . scc`, with a `fromJust` behind the...

Although heavyweight, phantom type parameters can bring other benefits too. We could use them for capturing all these graph variations in a uniform way: * Acyclic vs cyclic * Nonempty...

@subttle I'm not sure how linear types can help. Could you clarify?

@chessai Thanks, indeed! Also, `massiv` is close in spirit: http://hackage.haskell.org/package/massiv-0.2.6.0/docs/Data-Massiv-Array.html

@nobrakal Many thanks for the experiment! I have been wondering recently: do we even need graphs that can be empty? Presumably, non-empty graphs cover 99% of use-cases and are generally...

> I don't understand yet too... It is a huge drop for a single comparison ! Perhaps, it prevents some important optimisation? Could you investigate by looking at generated Core?

@nobrakal Why not use `edges1` in the second case? Then you don't need to reimplement it.

> note that the rough number for `hasVertex` is a false positive Hmm, this makes me think that the benchmarking suite is too fragile: it shouldn't measure performance by looking...