Make kernel laws consistent with core laws by LukaJCB · Pull Request #1922 · typelevel/cats

LukaJCB · 2017-09-20T15:59:54Z

Should fix #1272, still a WIP, but compiles and runs so looking for some feedback on some of the decisions made when porting between styles :)

codecov-io · 2017-09-21T08:35:19Z

Codecov Report

Merging #1922 into master will increase coverage by 0.53%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #1922      +/-   ##
==========================================
+ Coverage    95.6%   96.13%   +0.53%     
==========================================
  Files         252      273      +21     
  Lines        4570     4530      -40     
  Branches      124      116       -8     
==========================================
- Hits         4369     4355      -14     
+ Misses        201      175      -26

Impacted Files	Coverage Δ
...main/scala/cats/kernel/laws/SerializableLaws.scala	`93.75% <ø> (ø)`
.../src/main/scala/cats/kernel/instances/option.scala	`100% <ø> (ø)`	⬆️
...ats/kernel/laws/discipline/SerializableTests.scala	`100% <ø> (ø)`
...ats/kernel/laws/discipline/SemigroupLawTests.scala	`100% <100%> (ø)`
.../scala/cats/kernel/laws/discipline/BandTests.scala	`100% <100%> (ø)`
...cats/kernel/laws/discipline/SemilatticeTests.scala	`100% <100%> (ø)`
...la/cats/kernel/laws/discipline/OrderLawTests.scala	`100% <100%> (ø)`
...ws/src/main/scala/cats/kernel/laws/OrderLaws.scala	`100% <100%> (ø)`	⬆️
laws/src/main/scala/cats/laws/package.scala	`100% <100%> (ø)`	⬆️
.../src/main/scala/cats/laws/discipline/package.scala	`100% <100%> (ø)`	⬆️
... and 54 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 4ffd3a1...668c4e1. Read the comment docs.

LukaJCB · 2017-09-22T10:45:32Z

I moved most of the machinery from cats.laws.* to cats.kernel.laws.*, including IsEq, SerializableLaws and SerializableTests. By doing so, I removed the seperate IsSerializable object, as I considered the test inside to be equivalent to the one inside SerializableLaws (Please correct me here if I'm wrong)

Otherwise I'm pretty much done with this :)

johnynek · 2017-09-22T17:28:30Z

note algebra-laws do depend on this, so we will need to fix algebra before it can be updated after this is published:

https://github.com/typelevel/algebra/blob/master/build.sbt#L143

LukaJCB · 2017-09-23T08:41:47Z

I'll create a PR :)

kailuowang · 2017-09-27T18:22:51Z

+  def laws: BoundedSemilatticeLaws[A]
+
+  def boundedSemilattice(implicit arbA: Arbitrary[A], eqA: Eq[A]): RuleSet =
+    new RuleSet {


nit: most of these new RuleSet { where bases is empty could be just new DefaultRuleSet( right like the one you used in BandTests? No big deal though, just that it could be slightly more concise.

That doesn't work for Typeclasses with more than one parent though, no?

I didn't realize that. disregard my comment then. Maybe we can create one that does in discipline later.

kailuowang · 2017-09-27T18:38:35Z

-import org.scalacheck.{Arbitrary, Prop}
-import cats.kernel.instances.boolean._
-
-object Rules {


This setup is indeed less readable than the cats.core way.

kailuowang · 2017-09-27T18:48:22Z

 package tests

-import cats.kernel.laws.GroupLaws
+import cats.kernel.laws.discipline.{MonoidTests => MonoidTypeclassTests}


Another nit: Maybe we can call these MonoidLawTests, because the other MonoidTest is also testing the Type Class.

kailuowang · 2017-09-27T18:58:29Z

+
+import cats.kernel.Monoid
+
+trait MonoidLaws[A] extends SemigroupLaws[A] {


I couldn't find the replacement for Rules.isId("isEmpty", A.empty)(A.isEmpty)?

kailuowang · 2017-09-27T20:06:57Z

+    new RuleSet {
+      val name: String = "boundedSemilattice"
+      val bases: Seq[(String, RuleSet)] = Nil
+      val parents: Seq[RuleSet] = Seq(commutativeSemigroup, band)


should this be communtativeMonoid, semilattice instead?

kailuowang · 2017-09-28T16:52:45Z

+  override implicit def E: PartialOrder[A]
+
+  def reflexitivity(x: A): IsEq[Boolean] =
+    E.gteqv(x, x) <-> true


the original test uses ?<=, i.e. E.lteqv, I am not sure if it matters. (mabye we can have both lteqv and gteqv?)

kailuowang · 2017-10-04T15:19:23Z

  checkAll("Function0[Eqed]", EqLawTests[Function0[Eqed]].eqv)
  checkAll("Function0[POrd]", PartialOrderLawTests[Function0[POrd]].partialOrder)
  checkAll("Function0[Ord]", OrderLawTests[Function0[Ord]].order)
-  checkAll("Function0[Hsh]", HashLaws[Function0[Hsh]].hash)


Hmm, was this test removed intentionally?

Good catch, I removed it initially for erroring for erroring out, as with the new setup, all the laws are tested and you can't single out one specific law to run. I wanted to replace it with a manual test, but forgot, will fix :)

LukaJCB · 2017-10-04T17:29:09Z

I added another MiMa exception as I had to reshuffle the option instances :)

kailuowang · 2017-10-05T02:18:34Z

  checkAll("Function0[Grp]", GroupLawTests[Function0[Grp]].group)
  checkAll("Function0[CGrp]", CommutativeGroupTests[Function0[CGrp]].commutativeGroup)

+  test("Function0[Hsh]") {


just curious why not just test the hash law? won't pass?

kailuowang · 2017-10-05T02:19:34Z

+  checkAll("Function0[Grp]", GroupLawTests[Function0[Grp]].group)
+  checkAll("Function0[CGrp]", CommutativeGroupTests[Function0[CGrp]].commutativeGroup)
+
+  test("Function0[Hsh]") {


Just curious, why not just test the full HashLaw? won't pass?

Yeah, it's not the same as universal hash and not the same as scala.util.Hashing

kailuowang · 2017-10-05T02:21:59Z

 package object option extends OptionInstances

-trait OptionInstances extends OptionInstances1 {
+trait OptionInstances extends OptionInstances2 {


since we are reshuffling the number here, how about the we fix the priority number in one go? should be in the ascending order. BTW, @johnynek is this break okay?

I renamed the numbers :) Btw, I think this break is an oversight from the Hash PR, as PartialOrder and Hash instances had the same priority and both extend Eq, so it led to ambiguous implicits when requiring Eq for Option

kailuowang · 2017-10-05T12:49:15Z

filter with: ProblemFilters.exclude[UpdateForwarderBodyProblem]("cats.kernel.instances.OptionInstances0.catsKernelStdEqForOption")
filter with: ProblemFilters.exclude[ReversedMissingMethodProblem]("cats.kernel.instances.OptionInstances0.catsKernelStdPartialOrderForOption")
filter with: ProblemFilters.exclude[InheritedNewAbstractMethodProblem]("cats.kernel.instances.OptionInstances1.catsKernelStdHashForOption")
 filter with: ProblemFilters.exclude[MissingTypesProblem]("cats.kernel.instances.OptionInstances1")

LukaJCB · 2017-10-05T13:19:45Z

Thanks @kailuowang! Fixed it :)

kailuowang · 2017-10-06T11:46:58Z

sorry, more merge needed.

LukaJCB · 2017-10-06T12:12:05Z

No worries, done :)

johnynek

Thanks for doing this huge job. This is a great service to cats and I appreciate it.

johnynek · 2017-09-22T17:40:09Z

- * Object with a dynamic variable that allows users to skip the
- * serialization tests for certain instances.
- */
-private[laws] object IsSerializable {


did we lose this test entirely?

It is very important for spark and scalding (and others).

No, it's replaced by the SerializableLaws moved from cats.tests

adelbertc · 2017-10-06T19:41:49Z

Oh damn just saw this @LukaJCB - you the man!

LukaJCB force-pushed the make-kernel-laws-consistent branch 2 times, most recently from d66e8a7 to 29f9434 Compare September 21, 2017 07:25

LukaJCB force-pushed the make-kernel-laws-consistent branch 3 times, most recently from 7785850 to 0616d9b Compare September 22, 2017 09:58

kailuowang mentioned this pull request Sep 27, 2017

Add Invariant instances for kernel type classes #1914

Merged

kailuowang reviewed Sep 27, 2017

View reviewed changes

LukaJCB force-pushed the make-kernel-laws-consistent branch from 55c82a5 to 881553e Compare September 28, 2017 06:01

Luka Jacobowitz added 12 commits September 28, 2017 09:05

Make GroupLaws consistent with cats-core

20806ee

Make OrderLaws consistent with core

269f0e9

Port tests

ea453b3

Update lawtesting docs

6bebba6

Remove unused machinery

b9b2264

Remove duplication of IsEq

28b0b41

Cleanup

b1fb806

Add Serializable to LawTests

e50bf06

Rename to -LawTests

b57dc31

Fix BoundedSemilattice parents

b6de9dd

Add id

f81d7b2

Fix lawtesting docs

5073a95

LukaJCB force-pushed the make-kernel-laws-consistent branch from 881553e to 5073a95 Compare September 28, 2017 07:08

kailuowang reviewed Sep 28, 2017

View reviewed changes

kailuowang previously approved these changes Oct 4, 2017

View reviewed changes

LukaJCB dismissed kailuowang’s stale review via 15995be October 4, 2017 15:05

LukaJCB force-pushed the make-kernel-laws-consistent branch from 15995be to d3464b1 Compare October 4, 2017 15:07

kailuowang reviewed Oct 4, 2017

View reviewed changes

Convert HashLaws to new setup

8dbbb81

LukaJCB force-pushed the make-kernel-laws-consistent branch from d3464b1 to 8dbbb81 Compare October 4, 2017 15:58

Luka Jacobowitz added 2 commits October 4, 2017 18:25

Readd function0 hash test

ca30cd4

Add mima exception

08e1bcd

kailuowang reviewed Oct 5, 2017

View reviewed changes

LukaJCB force-pushed the make-kernel-laws-consistent branch from 2a33438 to d82b447 Compare October 5, 2017 13:19

LukaJCB force-pushed the make-kernel-laws-consistent branch from d82b447 to fb163f3 Compare October 5, 2017 13:58

Rename to ascending order

1ca0ecd

LukaJCB force-pushed the make-kernel-laws-consistent branch from fb163f3 to 1ca0ecd Compare October 5, 2017 13:58

kailuowang previously approved these changes Oct 5, 2017

View reviewed changes

Merge branch 'master' into make-kernel-laws-consistent

668c4e1

LukaJCB dismissed kailuowang’s stale review via 668c4e1 October 6, 2017 12:06

kailuowang approved these changes Oct 6, 2017

View reviewed changes

johnynek approved these changes Oct 6, 2017

View reviewed changes

kailuowang merged commit 9ae4dc4 into typelevel:master Oct 6, 2017

kailuowang added this to the 1.0.0-RC1 milestone Oct 13, 2017

kailuowang added the testing label Oct 15, 2017

johnynek mentioned this pull request Nov 3, 2017

Update Cats Version and Other Compatible Shifts typelevel/algebra#204

Merged


		import cats.kernel.Monoid

		trait MonoidLaws[A] extends SemigroupLaws[A] {

Uh oh!

Uh oh!

Conversation

LukaJCB commented Sep 20, 2017

Uh oh!

codecov-io commented Sep 21, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

LukaJCB commented Sep 22, 2017

Uh oh!

johnynek commented Sep 22, 2017

Uh oh!

LukaJCB commented Sep 23, 2017

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

LukaJCB commented Oct 4, 2017

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

kailuowang commented Oct 5, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

LukaJCB commented Oct 5, 2017

Uh oh!

kailuowang commented Oct 6, 2017

Uh oh!

LukaJCB commented Oct 6, 2017

Uh oh!

johnynek left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

adelbertc commented Oct 6, 2017

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

codecov-io commented Sep 21, 2017 •

edited

Loading

kailuowang commented Oct 5, 2017 •

edited

Loading