Add IdT, the identity monad transformer

[vikraman]

No description provided.

[vikraman]

I don't see a MonadTrans yet, should that be added?

[codecov-io]

Current coverage is 84.00%

Merging #588 into master will decrease coverage by -0.15% as of 15596c2

@@            master   #588   diff @@
=====================================
  Files          162    163     +1
  Stmts         2222   2244    +22
  Branches        74     74       
  Methods          0      0       
=====================================
+ Hit           1870   1885    +15
  Partial          0      0       
- Missed         352    359     +7

Review entire Coverage Diff as of 15596c2

Powered by Codecov. Updated on successful CI builds.

[ceedubs]

Thanks, @vikraman!

I've never used IdT before. Do you have an example of how/why one might want to use it?

[xuwei-k]

I think IdT should be an own data type, not type alias.

[non]

@vikraman What do you think? Would you be willing to make this its own data type as @xuwei-k suggests? I am starting to feel like we should avoid type aliases in cases like these.

[vikraman]

Yes, I just ran into diverging implicits with this alias, I'll change it to a proper data type.

[vikraman]

Updated.

[aryairani]

Ditto to @ceedubs's question

[julien-truffaut]

should be replace the constraint from Monad to FlatMap?

[vikraman]

I think it is useful to have IdT since it is the most trivial transformer. It would be useful to have the definitions of liftM once there is a MonadTrans class.

[xuwei-k]

xuwei-k@5b426cc249bd1f46d92b8d2

[error] /home/travis/build/xuwei-k/cats/tests/src/test/scala/cats/tests/IdTTests.scala:10: ambiguous implicit values:
[error]  both method idTFunctor in class IdTInstances of type [F[_]](implicit F: cats.Functor[F])cats.Functor[[β]cats.data.IdT[F,β]]
[error]  and method idTMonad in class IdTInstances of type [F[_]](implicit F: cats.Monad[F])cats.Monad[[β]cats.data.IdT[F,β]]
[error]  match expected type cats.Functor[[β]cats.data.IdT[F,β]]
[error]   def idTFunctor[F[_]: Monad] = implicitly[Functor[IdT[F, ?]]]
[error]                                           ^

[vikraman]

Split the instances to avoid ambiguous implicits.

[non]

Sorry for the long silence.

I'm fine with merging this -- I do think we probably want to write a few more words about how it works (e.g. "If M[_] is a monad, then IdT[M, ?] provides an equivalent monad" or similar) but that can happen in a future PR.

👍

[julien-truffaut]

LGTM 👍