Support OrganizeImports.removeUnused in Scala 3.4+

[tanishiking]

Second half of / closes #1682

This PR is mostly for verify scala/scala3#17835 actually unlock OrganizeImports.removeUnused (and UnusedImports) in scalafix, and it looks like it does 🎉

Screen.Recording.2023-06-30.at.14.57.19.mov

Once this PR scala/scala3#17835 has merged and released, scalafix-organize-imports should be able to run OrganizeImports.removeUnused based on the diagnostics information in SemanticDB emit from Scala3 compiler.

In order to make OrganizeImports rule to work with Scala 3, this PR added a few adjustments to the rule.

[tanishiking]

RemoveUnused can be supported in the same way #1728 👍

[bjaglin]

Thank you so much for your work upstream and taking the time to double-check it here! Added some comments for later, I'll be happy to take over if you don't have bandwidth.

Doc updates I have identified so far:

• scalafix/docs/rules/OrganizeImports.md

  Lines 55 to 57 in 0cf7bd7

  	1. The [`removeUnused`](OrganizeImports.md#removeunused) option must be
  	explicitly set to `false` - the rule currently doesn’t remove unused
  	imports as it is currently not supported by the compiler.

• scalafix/docs/rules/OrganizeImports.md

  Lines 1282 to 1287 in 0cf7bd7

  	> The `removeUnused` option is currently not supported for source files
  	> compiled with Scala 3, as the [compiler cannot issue warnings for unused
  	> imports
  	> yet](https://docs.scala-lang.org/scala3/guides/migration/options-lookup.html#warning-settings).
  	> As a result, you must set `removeUnused` to `false` when running the
  	> rule on source files compiled with Scala 3.

• scalafix/docs/rules/OrganizeImports.md

  Line 1302 in 0cf7bd7

  removeUnused = true // not supported in Scala 3

[tanishiking]

Thank you for the comments! I'll get back to this once the upstream change has merged, maybe I'll ask someone to take over (maybe I'll do by myself). Anyway, I'll let you know :)

[tanishiking]

Finally scala/scala3#17835 is merged, and we'll be able to support this feature from the next release of the Scala3 🎉

[tanishiking]

I think we don't need to rush until 3.4.0 is released, but now the test works fine with

$ sbt -Dscala3.nightly=3.4.0-RC1-bin-20231024-15033c7-NIGHTLY
sbt:scalafix> expect3Target3_4_0-RC1-bin-20231024-15033c7-NIGHTLY/test

[bjaglin]

I think we don't need to rush until 3.4.0 is released, but now the test works fine with

For the inpatients (like me!) wondering about the timing, see scala/scala3#17835 (comment)

[bjaglin]

I've got #1728 sorted out and there are conflicts, so I'll squash and rebase your PR.

[bjaglin]

https://github.com/scalacenter/scalafix/actions/runs/7767691536/job/21184796798?pr=1800

[error] --- /home/runner/work/scalafix/scalafix/scalafix-tests/integration/src/main/scala/scalafix/test/cli/Rules.scala
[error] +++ <expected fix>
[error] @@ -1,7 +1,5 @@
[error]  package scalafix.test.cli
[error]  
[error] -import scala.meta._
[error] -
[error]  import metaconfig.Configured
[error]  import scalafix.internal.util.PositionSyntax._
[error]  import scalafix.internal.util.SymbolOps
...
[error] (integration2_13 / scalafixAll) scalafix.sbt.ScalafixFailed: TestError

But the file won't compile without that...

[bjaglin]

This is clearly a side effect of this PR as we are dogfooding, but it might be specific to the complex build of the project and not representative. The weird thing is that I can't reproduce locally. Investigating.

[bjaglin]

Relaxing the isUnused() implementation exposed a pretty serious bug: the mutable collections used in OrganizeImports are not cleared from one file to another, as the class is instantiated only once per scalafix invocation.

scalafix/scalafix-rules/src/main/scala/scalafix/internal/rule/OrganizeImports.scala

Lines 65 to 69 in 9271030

	private val unusedImporteePositions: mutable.Set[Position] =
	mutable.Set.empty[Position]
	private val diagnostics: ArrayBuffer[Diagnostic] =
	ArrayBuffer.empty[Diagnostic]

I'll fix it in a separate PR, and rebase this PR to remove the need for the hacky second commit.

[bjaglin]

Rebased against #1921. This is it ✌️ I can't believe this is finally happening 🥳

Thanks so much for the perseverance on the dotty front @tanishiking !