Lint dubious overload differs only in implicit [ci: last-only] #11033

som-snytt · 2025-03-31T00:42:33Z

~~Tweaks~~ Expands upon the truly retro (and pithy) retronym WIP.

lrytz

LGTM

test/files/neg/t7415.scala

src/compiler/scala/tools/nsc/typechecker/RefChecks.scala

som-snytt · 2025-03-31T20:34:13Z

The other possible advice is to ensure the implicit param has implicitNotFound message to explain what happens when it doesn't compile as expected.

som-snytt · 2025-04-04T01:23:53Z

This version takes nilary and implicit members, group by name for comparison, then looks at groups of more than one such-named members.

Sorry, @lrytz

lrytz · 2025-04-04T08:52:00Z

Didn't review yet, but the current version has an issue

[error] /home/jenkins/workspace/scala-2.13.x-validate-main/src/library/scala/collection/immutable/NumericRange.scala:492:40: Usages of value step will be easy to mistake for calls to private[this] val step: T, which has a single implicit parameter list.
[error]   class Inclusive[T](start: T, end: T, step: T)(implicit num: Integral[T])
[error]                                        ^

som-snytt · 2025-04-04T12:05:20Z

Sorry, I have another local commit to do accessors correctly. The desideratum is to warn for the private local in the test. Oh I see I was in the middle of an edit yesterday.

It warns for jdk.DoubleAccumulator which has toArray and toArray[B: ClassTag]. Hmm.

def toArray: Array[Double]
def toArray[B >: Double](implicit evidence$2: scala.reflect.ClassTag[B]): Array[B]

scala> acc.toArray

I guess this is what nowarn is for.

som-snytt · 2025-04-04T15:42:28Z

test/files/neg/t7415.check

+            ^
+t7415.scala:66: warning: Usages of value i will be easy to mistake for calls to def i(implicit t: T): Int, which has a single implicit parameter list.
+class S(i: Int) extends R(i) { // warn
+      ^


Could tweak the message to say Usages of inherited member... to clarify. The caret is at the class, but that is subtle. Cf new Derived1 {} below

I did not follow up, it was too subtle for me.

som-snytt · 2025-04-04T15:56:12Z

src/library/scala/jdk/DoubleAccumulator.scala

@@ -236,6 +237,7 @@ final class DoubleAccumulator
  }

  /** Copies the elements in this `DoubleAccumulator` into an `Array[Double]` */
+  @nowarn // cat=lint-overload see toArray[B: ClassTag]
  def toArray: Array[Double] = {


This suggests a limitation of "manual specialization", namely the change in signature.

Would I ever want to acc.toArray[Any]?

Not sure yet whether to suppress for this case? or is this exactly what the lint is for?

lrytz · 2025-04-14T08:57:00Z

src/compiler/scala/tools/nsc/typechecker/RefChecks.scala

+      def ofInterest(tp: Type): Boolean = tp match {
+        case mt: MethodType => mt.isImplicit
+        case PolyType(_, rt) => ofInterest(rt)
+        case _: NullaryMethodType | _ => true


Suggested change

case _: NullaryMethodType | _ => true

case _ => true

_: NullaryMethodType | _ could be considered clarification or confusing, depending on the time of day..

Normally I'd put it in /* _: NullaryMethodType |*/ and then a reviewer would say delete?. I guess it could go in a regular line comment. This is an example where one forgets the model (that there is a nullary MT) if one has been away from it. Why doesn't the compiler help? I don't want to enumerate all the types I'm not interested in.

lrytz · 2025-04-14T08:57:14Z

src/compiler/scala/tools/nsc/typechecker/RefChecks.scala

+      def isNilary(tp: Type): Boolean = tp match {
+        case _: MethodType => false
+        case PolyType(_, rt) => isNilary(rt)
+        case _: NullaryMethodType | _ => true


Suggested change

case _: NullaryMethodType | _ => true

case _ => true

lrytz · 2025-04-14T09:12:34Z

src/compiler/scala/tools/nsc/typechecker/RefChecks.scala

+        case PolyType(_, rt) => ofInterest(rt)
+        case _: NullaryMethodType | _ => true
+      }
+      def isNilary(tp: Type): Boolean = tp match {


Didn't we call foo "nullary" and foo() "nilary"? So this should be isNullary?

Maybe you're right. I thought nilary means paramss is Nil but maybe it's that it's ListOfNil. Probably I just forgot.

The evil comment on NullaryMethodType must be wrong then. What it says we mean by "nullary" is what we call "nilary". Then the comment in Namers in overridesNilary is correct: it checks for MethodType not NMT because it wants "nilary" (neither nullary nor nelary).

som-snytt · 2025-04-14T13:33:38Z

src/compiler/scala/tools/nsc/typechecker/RefChecks.scala

@@ -160,6 +160,72 @@ abstract class RefChecks extends Transform {
        })
      }
    }
+    private def checkDubiousOverloads(clazz: Symbol): Unit = if (settings.warnDubiousOverload) {
+      // nilary members or methods with leading implicit params


so this comment should say nullary. Just to confirm, we're saying the NullaryMethodType is correctly named, and "nilary" was coined so we wouldn't have to rename it.

som-snytt · 2025-04-14T13:48:36Z

Thanks, I fixed what would have been a terrible faux pas.

lrytz

Nice, thank you!

scala-jenkins added this to the 2.13.17 milestone Mar 31, 2025

som-snytt force-pushed the issue/7415-similar-overloads branch from 1d2e2ba to fc4e29e Compare March 31, 2025 00:53

som-snytt marked this pull request as ready for review March 31, 2025 05:21

som-snytt mentioned this pull request Mar 31, 2025

Find more doc vars and distinguish C from CC [ci: last-only] #10218

Merged

lrytz approved these changes Mar 31, 2025

View reviewed changes

test/files/neg/t7415.scala Show resolved Hide resolved

src/compiler/scala/tools/nsc/typechecker/RefChecks.scala Outdated Show resolved Hide resolved

som-snytt marked this pull request as draft March 31, 2025 16:09

som-snytt changed the title ~~Lint dubious overload differs only in implicit~~ Lint dubious overload differs only in implicit [ci: last-only] Apr 4, 2025

som-snytt force-pushed the issue/7415-similar-overloads branch from fc4e29e to 09eda7c Compare April 4, 2025 01:17

som-snytt marked this pull request as ready for review April 4, 2025 01:24

som-snytt requested a review from lrytz April 4, 2025 01:24

som-snytt marked this pull request as draft April 4, 2025 12:01

som-snytt force-pushed the issue/7415-similar-overloads branch from c754eb2 to a362aa7 Compare April 4, 2025 15:35

som-snytt commented Apr 4, 2025

View reviewed changes

som-snytt marked this pull request as ready for review April 4, 2025 18:10

lrytz reviewed Apr 14, 2025

View reviewed changes

som-snytt commented Apr 14, 2025

View reviewed changes

som-snytt force-pushed the issue/7415-similar-overloads branch from 4ed5e86 to 9c6a721 Compare April 14, 2025 13:46

Lint dubious overload differs only in implicit

d856b48

som-snytt force-pushed the issue/7415-similar-overloads branch from 9c6a721 to d856b48 Compare April 15, 2025 01:37

lrytz approved these changes Apr 15, 2025

View reviewed changes

lrytz merged commit 38cbd77 into scala:2.13.x Apr 15, 2025
3 checks passed

som-snytt deleted the issue/7415-similar-overloads branch April 15, 2025 13:53

xuwei-k mentioned this pull request Apr 29, 2025

Usages of value a will be easy to mistake for calls warning since Scala 2.13.17 scalatest/scalatest#2364

Open

Lint dubious overload differs only in implicit [ci: last-only] #11033

Lint dubious overload differs only in implicit [ci: last-only] #11033

Conversation

som-snytt commented Mar 31, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

lrytz left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

som-snytt commented Mar 31, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

som-snytt commented Apr 4, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

lrytz commented Apr 4, 2025

Uh oh!

som-snytt commented Apr 4, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

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!

som-snytt commented Apr 14, 2025

Uh oh!

lrytz left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

som-snytt commented Mar 31, 2025 •

edited

Loading

som-snytt commented Mar 31, 2025 •

edited

Loading

som-snytt commented Apr 4, 2025 •

edited

Loading

som-snytt commented Apr 4, 2025 •

edited

Loading