improve interaction with subtyping in OptionOps#liftTo

[mberndt123]

Currently OptionOps#liftTo interacts badly with subtyping. Something like this won't compile:

42.some.liftTo[IO](new Exception("foo"))

One needs to write this instead:

42.some.liftTo[IO](new Exception("foo"): Throwable)

This patch fixes this.

[codecov-io]

Codecov Report

Merging #2480 into master will not change coverage.
The diff coverage is 100%.

@@           Coverage Diff           @@
##           master    #2480   +/-   ##
=======================================
  Coverage   95.38%   95.38%           
=======================================
  Files         357      357           
  Lines        6517     6517           
  Branches      279      282    +3     
=======================================
  Hits         6216     6216           
  Misses        301      301

Impacted Files	Coverage Δ
core/src/main/scala/cats/ApplicativeError.scala	100% <ø> (ø)	⬆️
core/src/main/scala/cats/syntax/option.scala	100% <ø> (ø)	⬆️
core/src/main/scala/cats/syntax/either.scala	99.21% <100%> (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 92236f8...53f7cbb. Read the comment docs.

[kailuowang]

This is nice. Would you add a doctest to verify the new ergonomics?

[kailuowang]

Also the liftTo on Either can be improved the same way right?

[kailuowang]

Thanks! BTW, in case you haven't used doctest before, here is an example of PR adding doctests, it would nice to have test make sure that this new interaction works as intended.

[mberndt123]

I added the same change to EitherOps#liftTo.

I also added doctests. The latter turned out to be a PITA. I added the doctests, and they compiled and ran. In order to see whether they actually exercise my new code, I changed the _ >: E in OptionOps.LiftToPartiallyApplied part back to E and expected them to fail. But they still ran just fine, and I couldn't figure out what the hell was going on… Until, after quite a bit of head scratching, I ran clean in sbt before coreJVM/test and, lo and behold, the tests started failing as expected! Apparently incremental compilation with sbt/zinc/scalac is still broken ☹️

Oh well, one more scala war story to tell.

[LukaJCB]

Looks great, thank you!

[kailuowang]

This is nice. Thanks!

[mberndt123]

I've simplified the tests a bit. The reason I used EitherT[Option, CharSequence, ?] instead of just Either[CharSequence, ?] in the test is that my test was passing when I expected it to fail. I suspected that Either's covariance might have something to do with it, so I changed the test to use EitherT, which is invariant. But the actual cause turned out to be the above-mentioned incremental compilation bug, so there's no reason to complicate things with EitherT.

[mberndt123]

@kailuowang @LukaJCB could you approve again?

[mberndt123]

This PR was approved and all review comments were addressed, so to me it looks like it's ready to be merged. Is there anything else to do to get this merged?

[LukaJCB]

Sorry for the delay, merging now :)

[mberndt123]

Cool, thanks!