Add a few type classes to generated tuple instances

[edmundnoble]

No description provided.

[peterneyens]

Should we add instances for CommutativeSemigroup, CommutativeGroup and BoundedSemilattice while we are at it?

[johnynek]

We're going to need generators for our generators.

Are we getting the priorities correct? We want CommutativeX to be a higher priority than X.

It would be interesting if we had some better way to generalize over this.

[johnynek]

I often wonder how long this approach is the right one and whether a shapeless based approach would be almost as fast and much cleaner to maintain.

[codecov-io]

Codecov Report

Merging #1527 into master will decrease coverage by 2.2%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #1527      +/-   ##
==========================================
- Coverage   94.02%   91.82%   -2.21%     
==========================================
  Files         256      246      -10     
  Lines        4201     3754     -447     
  Branches       84      134      +50     
==========================================
- Hits         3950     3447     -503     
- Misses        251      307      +56

Impacted Files	Coverage Δ
core/src/main/scala/cats/syntax/reducible.scala	40% <0%> (-60%)	⬇️
core/src/main/scala/cats/Monad.scala	50% <0%> (-46.16%)	⬇️
core/src/main/scala/cats/Apply.scala	55.55% <0%> (-33.34%)	⬇️
laws/src/main/scala/cats/laws/CategoryLaws.scala	66.66% <0%> (-33.34%)	⬇️
core/src/main/scala/cats/MonadWriter.scala	0% <0%> (-33.34%)	⬇️
core/src/main/scala/cats/data/package.scala	66.66% <0%> (-19.05%)	⬇️
core/src/main/scala/cats/instances/try.scala	78.94% <0%> (-17.42%)	⬇️
...ain/scala/cats/laws/discipline/CategoryTests.scala	83.33% <0%> (-16.67%)	⬇️
core/src/main/scala/cats/syntax/either.scala	83.92% <0%> (-15.21%)	⬇️
core/src/main/scala/cats/data/Cokleisli.scala	88.88% <0%> (-11.12%)	⬇️
... and 114 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 bbc97b8...7eb3e8d. Read the comment docs.

[edmundnoble]

@johnynek What about Monoid and Semigroup for example, shouldn't they be different priorities too?

@peterneyens I added the three suggested as well.

Note this breaks bincompat, and so needs to be for the next release of cats-kernel, whenever that will be.

[kailuowang]

If there is concerns over breaking bin comp for kernel we can add these instances in cats.core right?

[johnynek]

these are fine from algebird's point of view. We don't use these traits directly.

[johnynek]

@edmundnoble yes, I think the priority issue might also be a problem for Semigroup. I think it was an oversight. We should probably have resolution tests to make sure things work as expected. Just using Tuple2 is probably enough.

[kailuowang]

@edmundnoble I wonder if you have time to finish up this PR? Looks like we are close (1. fix the priority, and 2. add mima exception)

[kailuowang]

Continued in #1967