Andrey Mokhov

Thanks @subttle! While the instances look reasonable, it's not obvious to me that Applicative and Monad laws hold. Could you give some intuition about why you chose this particular implementation,...

@subttle Strange: I received an email notification with your comment about using QuickCheck, but it doesn't seem to be here. Have you deleted it, or is it a GitHub glitch?...

Thanks @subttle! One simple thing to do to at least resolve the question about the validity of the Applicative instance is to use `() = ap` as the implementation. This...

@jmtd Cool, thank you! I like how nicely it all worked out. There seem to be a performance regression for `hasVertex` and possibly `hasEdge` too: ``` hasEdge: 1.17 (bad) hasVertex:...

Adding `INLINE` pragmas seems to have fixed the regression. (There is still this strange result `hasEdge: 0.69 (good)`, which we believe is a bug.) @jmtd Could you try the following?...

@jmtd Oh, I'm sorry, I meant: ```haskell findVertices p = foldg [] (\x -> [ x | p x ]) (++) (++) ```

@jmtd Thanks for giving it a try. I suggest we proceed as follows. Let's keep the `Algebra.Graph` module as is, but instead add new functionality to the generic module that...

@jmtd Thanks! Let me think about this for a few days. I'm travelling until 11 November, so it's a bit difficult to find time to figure out a small set...

@mateoatr Many thanks for the PR! I've added a couple of comments above. Also two general comments: * For some reason you do not cover the most important data type...

@jmtd Thank you! I just realised that I might have overcomplicated things by suggesting to add the type class, since we can escape after we have the `String`, just like...