filterM broken?

[3rdLaw]

I ran the following in sbt console after a git pull:

scala> import cats.implicits._
import cats.implicits._
scala> import cats._
import cats._
scala> List(1, 2, 3).filterM { x => List(true, false) }
res0: List[Int] = List(1, 2, 3)

Compare this to scalaz's results:

import scalaz._
import scalaz.Scalaz._
List(1, 2, 3).filterM { x => List(true, false) }
res0: List[List[Int]] = List(List(1, 2, 3), List(1, 2), List(1, 3), List(1), List(2, 3), List(2), List(3), List())

[ceedubs]

@3rdLaw thanks for bringing this up. The filterM in cats indeed isn't the filterM in Haskell or scalaz. It doesn't even have quite the same type signature: the one in cats is not specialized to List and returns an F[A] result as opposed to F[F[A]].

We should probably do something to avoid confusion. @non it looks like you originally added filterM - what are your thoughts?

[3rdLaw]

Thanks for the clarification @ceedubs; I clearly assumed this was the same as Haskell/Scalaz's, so some documentation would be helpful.

Does cats have an equivalent to filterM in Haskell or scalaz?

[ceedubs]

@non to avoid confusion, what do you think about renaming our fliterM to flatFilter to parallel the relationship between map and flatMap?

[ceedubs]

@non on second thought, can you think of any practical example of using the method as it currently exists in cats? Option is about the only type that I imagine myself using it with, and in that case it would be pretty easy to use flatMap with filter instead.