Unactionable warning: "Given search preference for ... will change."

[eed3si9n]

Compiler version

3.6.4-RC1-bin-20241204-a67dbb0-NIGHTLY

Minimized code

You can grab sbt/sbt develop branch fdc03b782a40ce1f04bcefcd7eb3dc45fa7927d0.

I'm guessing that the minimized repro is basically https://github.com/scala/scala3/blob/c61897d7e78bfea283f5a2e37215c3b528e7684f/tests/warn/i21036a.scala

trait A
trait B extends A
given b: B = ???
given a: A = ???

val y = summon[A] // warn

Output

Currently the compile fails because I have -Xfatal-warnings and there's a warning about a given search:

[warn] -- Warning: /sbt/protocol/src/main/contraband-scala/sbt/protocol/codec/SettingQuerySuccessFormats.scala:14:91
[warn] 14 |      val value = unbuilder.readField[sjsonnew.shaded.scalajson.ast.unsafe.JValue]("value")
[warn]    |                                                                                           ^
[warn]    |Given search preference for sjsonnew.JsonReader[sjsonnew.shaded.scalajson.ast.unsafe.JValue] between alternatives
[warn]    |  (SettingQuerySuccessFormats.this.JValueFormat :
[warn]    |  sjsonnew.JsonFormat[sjsonnew.shaded.scalajson.ast.unsafe.JValue])
[warn]    |and
[warn]    |  (SettingQuerySuccessFormats.this.JValueJsonReader :
[warn]    |  sjsonnew.JsonReader[sjsonnew.shaded.scalajson.ast.unsafe.JValue])
[warn]    |will change.
[warn]    |Current choice           : the first alternative
[warn]    |New choice from Scala 3.7: the second alternative

Expectation

I expect all warnings to be actionable. This type of "FYI" message should go into the release note or -Xmigration 3.5.0.

[som-snytt]

I don't juggle implicits, but I appreciate the need to make warnings progressive and useful.

I think this warning has the force of deprecation.

In this case, trivially, it should summon[B] to maintain behavior. This is similar to the change in inference of overriding method types, where you add the explicit type to pick a behavior. It is inconvenient noise, but not inactionable.

"I spent hours getting my house of implicit cards just right and you broke it" is more annoying.

Lukas invented cat=scala3-migration to cope more conveniently with migration noise. I remember categories are not as leveraged in dotty.

m-w.com has synonyms for actionable: functional, applicative, pragmatic.
and antonyms: academic, useless, armchair.
I chose to say inactionable because inaction (intentional or not) is the result.

[eed3si9n]

m-w.com has synonyms for actionable: functional, applicative, pragmatic

Yes, in functional programming it's entirely conceivable, even pragmatic to define a tower of typeclasses like functor, apply, applicable, and monad. The change to summon the closest type like Applicative[List] as opposed to Monad[List] sounds great to me.

In this case, trivially, it should summon[B] to maintain behavior.

At least in this context, of JsonReader vs JsonFormat I'm not too concerned which one gets picked, but what I do care about is summoning the right typeclass, in this case JsonReader.

I chose to say inactionable because inaction (intentional or not) is the result.

I'd love to not take action, but with -Xfatal-warning, I am forced to take some action, so it's not even inactionable.

I remember categories are not as leveraged in dotty.

I just discovered https://scala-lang.org/2024/08/19/given-priority-change-3.7.html from reading https://scala-lang.org/news/3.6.2, and it seems like I can take one of the tree actions:

1. -source:3.5
2. -source:3.7
3. @nowarn("msg=Given search preference")

If we put in numbers to error messages, we could do @nowarn("id=E205"), which might be better than category.

[eed3si9n]

Here's my PR for this - #22189