standardise on liftF and add liftK to transformers

[SystemFw]

This PR deprecates lift on all transformers in favour of liftF, and adds a FunctionK version named liftK (for use with mapK).

I'm leaving a possible scalafix to a separate PR.

[kailuowang]

There probably won't be a rc2, the next one is 1.0

[SystemFw]

Yeah, wasn't quite sure. I'll fix tomorrow

[SystemFw]

Fixed the RC deprecation warning. I'm adding the scalafix now (changed my mind, I'll do it in this PR)

[kailuowang]

also need some mima exclusion

�[0m[�[31merror�[0m] �[0m filter with: ProblemFilters.excludeReversedMissingMethodProblem�[0m
�[0m[�[31merror�[0m] �[0m * method liftK(cats.Applicative,cats.kernel.Monoid)cats.arrow.FunctionK in trait cats.data.CommonIRWSTConstructors is present only in current version�[0m
�[0m[�[31merror�[0m] �[0m filter with: ProblemFilters.excludeReversedMissingMethodProblem�[0m
�[0m[�[31merror�[0m] �[0m * method liftF(java.lang.Object)cats.data.Kleisli in trait cats.data.KleisliFunctions is present only in current version�[0m
�[0m[�[31merror�[0m] �[0m filter with: ProblemFilters.excludeReversedMissingMethodProblem�[0m
�[0m[�[31merror�[0m] �[0m * method liftK()cats.arrow.FunctionK in trait cats.data.KleisliFunctions is present only in current version�[0m
�[0m[�[31merror�[0m] �[0m filter with: ProblemFilters.excludeReversedMissingMethodProblem�[0m
�[0m[�[31merror�[0m] �[0m * method liftF(java.lang.Object,cats.Applicative)cats.data.IndexedStateT in trait cats.data.CommonStateTConstructors is present only in current version�[0m
�[0m[�[31merror�[0m] �[0m filter with: ProblemFilters.excludeReversedMissingMethodProblem�[0m
�[0m[�[31merror�[0m] �[0m * method liftK(cats.Applicative)cats.arrow.FunctionK in trait cats.data.CommonStateTConstructors is present only in current version�[0m
�[0m[�[31merror�[0m] �[0m filter with: ProblemFilters.excludeReversedMissingMethodProblem�[0m

[SystemFw]

@kailuowang Any quick reference to how you exclude something from MiMa or should I RTFM? :)

[LukaJCB]

@SystemFw You can add them right here: https://github.com/typelevel/cats/blob/master/build.sbt#L212 :)

[SystemFw]

Thanks a lot! @LukaJCB

[aeons]

This is great. I've had to write the OptionT variant several times lately.

Is there a reason you're not using the kind-projector syntax?

def liftK[F[_]](implicit F: Functor[F]): F ~> OptionT[F, ?] = 
  λ[F ~> OptionT[F, ?]](OptionT.liftF(_))

(or Lambda instead of λ)

[SystemFw]

@aeons Yes, http4s was the initial use case :)

There's no reason for not using it, I might change it as I fix the other things

[SystemFw]

I think I'm mostly done, except my scalafix test doesn't compile ( it's telling me that RWST is not a member of cats.data ).
This happens because the scalafix is running on 0.9.0, and RWST was introduced later. How should I proceed?

P.S. the error in the travis seems spurious

[kailuowang]

the migration is for migrating from 0.9. ideally we need a new set of migration that migrates from 1.0 RC1, although I am not sure a simple rename in RWST justify that. I propose we just remove RWST from the test.

[codecov-io]

Codecov Report

Merging #2033 into master will decrease coverage by 0.17%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #2033      +/-   ##
==========================================
- Coverage   95.15%   94.98%   -0.18%     
==========================================
  Files         305      311       +6     
  Lines        5184     5267      +83     
  Branches      125      122       -3     
==========================================
+ Hits         4933     5003      +70     
- Misses        251      264      +13

Impacted Files	Coverage Δ
core/src/main/scala/cats/data/IndexedStateT.scala	98.8% <100%> (-1.2%)	⬇️
core/src/main/scala/cats/data/EitherT.scala	97.6% <100%> (+0.01%)	⬆️
core/src/main/scala/cats/data/Kleisli.scala	97.59% <100%> (-1.18%)	⬇️
core/src/main/scala/cats/data/WriterT.scala	93.75% <100%> (-0.94%)	⬇️
core/src/main/scala/cats/data/OptionT.scala	100% <100%> (ø)	⬆️
...in/scala/cats/data/IndexedReaderWriterStateT.scala	99.06% <100%> (-0.94%)	⬇️
core/src/main/scala/cats/Traverse.scala	100% <0%> (ø)	⬆️
core/src/main/scala/cats/Foldable.scala	100% <0%> (ø)	⬆️
...ain/scala/cats/laws/discipline/TraverseTests.scala	100% <0%> (ø)	⬆️
...a/cats/laws/discipline/NonEmptyTraverseTests.scala	100% <0%> (ø)	⬆️
... and 10 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 8cff547...e68cade. Read the comment docs.

[SystemFw]

@kailuowang I think the errors are spurious and I should be done. The scalafix has fixes for both 0.9 and 1.0RC, the RC ones are not tested as discussed above

[SystemFw]

Some of the builds fail with

Could not find any member to link for "mapK"

should I address this?

[LukaJCB]

That seems odd, seems like an error during scaladoc compilation. Maybe someone else has an idea?

[SystemFw]

Yeah, I reference [[mapK]] in a scaladoc, might be a pretty stupid error ? . Can't really test locally since MiMa fails before on my machine ("versions not set" so it's my setup, it succeeds on Travis)

[LukaJCB]

I think it might be because mapK is on the instance while liftF is on the companion object, so they need to be referred to differently. Not sure how exactly though.

[SystemFw]

I think we agree on what the problem is then :)
I also think we agree in not knowing what the solution is :P
I'll investigate (this sort of search and replace PRs always end up taking ages for me)

[kailuowang]

Linking reference in scaladoc has always been problematic, the solutions people have been doing in the past, and I propose doing here, is to just remove the link.

[SystemFw]

so just mapK rather than [[mapK]] ? Sounds good to me :)

[SystemFw]

I need to fix the doctest for kleisli. It's the only bit of codecov that actually spotted a problem. Doing so right now

[LukaJCB]

Thanks!

[frroliveira]

If you're still interested in getting the scaladoc references working, you can prefix mapK with the class name, e.g. EitherT.mapK. That's what I used in IorT.liftK

[kailuowang]

Close and reopen to trigger build