Issue 2657 - Ambiguous duration instances

[barambani]

This is to address #2657

[barambani]

I'm still trying to solve an issue here as importing both cats.instances.duration._ and cats.instances.finiteDuration._ explicitly prevents for the instance to be resolved. The scenario is when there's the need for kernel instances together with core's

object test {

  import cats.instances.duration._
  import cats.instances.finiteDuration._

  implicitly[Order[Duration]]

  implicitly[ContravariantShow[FiniteDuration]]
}

[kailuowang]

This is why we don't have variance in type classes. I'd say in this case user have to use workarounds.

[barambani]

I found a way to mitigate the problem. Now the only case that would fail is the below (see these test cases as an example).

test("fails") {
  import cats.instances.duration._
  import cats.instances.finiteDuration._

  implicitly[ContravariantShow[Duration]] // fails compilation
  implicitly[ContravariantShow[FiniteDuration]] // is inferred
}

I couldn't find a way to remove also this drawback.

[codecov-io]

Codecov Report

Merging #2670 into master will increase coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #2670      +/-   ##
==========================================
+ Coverage   95.12%   95.13%   +<.01%     
==========================================
  Files         364      365       +1     
  Lines        6707     6756      +49     
  Branches      285      300      +15     
==========================================
+ Hits         6380     6427      +47     
- Misses        327      329       +2

Impacted Files	Coverage Δ
testkit/src/main/scala/cats/tests/CatsSuite.scala	70% <ø> (ø)	⬆️
core/src/main/scala/cats/instances/duration.scala	100% <100%> (ø)	⬆️
...src/main/scala/cats/instances/finiteDuration.scala	100% <100%> (ø)	⬆️
core/src/main/scala/cats/syntax/parallel.scala	70.58% <0%> (-9.42%)	⬇️
...estkit/src/main/scala/cats/tests/ListWrapper.scala	100% <0%> (ø)

---

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 b329cf1...294b2e8. Read the comment docs.

[barambani]

I don't think I can do more here without breaking the BC. It's anyway a step forward compared to the current situation. If you agree, I would consider this PR complete.

[LukaJCB]

Thanks very much @barambani :)

[barambani]

👍

[kailuowang]

Thanks a lot! @barambani

[barambani]

👍