WIP: Show type class

[fthomas]

This is an attempt to allow for custom error messages (#161) and to improve string representations of compound predicates like type LessEqual[N] = Not[Greater[N]] (#49) using a dedicated Show type class.

First results look promising:

// using showExpr and showResult from Show

scala> refineV[LessEqual[_5]].apply2(10)
res1: Either[String,api.Refined[Int,numeric.LessEqual[shapeless.nat._5]]] =
  Left(Predicate failed: (10 <= 5).)

scala> refineV[NonEmpty].apply2("")
res2: Either[String,api.Refined[String,collection.NonEmpty]] =
  Left(Cannot create a NonEmptyString with the empty string)

// using showExpr and showResult from Validate

scala> refineV[LessEqual[_5]].apply(10)
res3: Either[String,api.Refined[Int,numeric.LessEqual[shapeless.nat._5]]] =
  Left(Predicate (10 > 5) did not fail.)

scala> refineV[NonEmpty].apply("")
res4: Either[String,api.Refined[String,collection.NonEmpty]] =
  Left(Predicate isEmpty() did not fail.)

[fthomas]

This could be also an opportunity to address #104.

[codecov[bot]]

Codecov Report

Merging #412 into master will decrease coverage by 2.34%. The diff coverage is 0%.

@@            Coverage Diff             @@
##           master     #412      +/-   ##
==========================================
- Coverage   95.51%   93.16%   -2.35%     
==========================================
  Files          46       47       +1     
  Lines         557      571      +14     
  Branches       12       10       -2     
==========================================
  Hits          532      532              
- Misses         25       39      +14

Impacted Files	Coverage Δ
...ed/src/main/scala/eu/timepit/refined/numeric.scala	64.28% <0%> (-4.95%)	:arrow_down:
...epit/refined/internal/RefinePartiallyApplied.scala	57.14% <0%> (-42.86%)	:arrow_down:
...d/src/main/scala/eu/timepit/refined/api/Show.scala	0% <0%> (ø)
...red/src/main/scala/eu/timepit/refined/string.scala	91.89% <0%> (-8.11%)	:arrow_down:

---

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 c223dab...a6a173b. Read the comment docs.

[kusamakura]

If we can separate out the result display part of Validate into this Show typeclass, I think the concern you raised in #397 can also be addressed cleanly.

[fthomas]

You're right @kusamakura. This is painful lesson in keeping type classes focused on one concern.

[satorg]

Hi guys. Just wondering – any activity on this PR? Seems this functionality is a really nice-to-have one.