(Scala 3) completions are more sensitive than Scala 2's

[kubukoz]

Describe the bug

In Scala 3 completions, it seems like whatever algorithm was used for fuzzy finding in Scala 2 has been changed to a more strict one. It's limiting the amount of completions that you get, and could be considered a defect in the functionality.

To reproduce:

//> using scala "2.13.10"

object MyInterface {
  def someMethod(s: String): Unit = ()
}

object demo {
  MyInterface.<CURSOR>
}

If you start typing at <CURSOR> and write m, you'll see a completion for the method:

In Scala 3, you don't see it:

Expected behavior

The method completion shows up.

Operating system

macOS

Editor/Extension

VS Code

Version of Metals

v0.11.9

Extra context or search terms

No response

[tgodzik]

Thanks for reporting! The issue is that the compiler is trying to filter the results first, we need to change that upstream I think.

[tgodzik]

We would need to change it around here https://github.com/lampepfl/dotty/blob/main/compiler/src/dotty/tools/dotc/interactive/Completion.scala#L136

[kubukoz]

nameInScope.startsWith(prefix) <- probably this part, right?

[tgodzik]

yep, Scala 2 allowed the downstream tools to use any kind of matcher