Ólafur Páll Geirsson

@wchau thank you for reporting! Can you upload the payload that’s failing to parse? This is the first time somebody reports this issue and SemanticDB sees quite heavy usage.

Would you be OK with privately sharing the payload with me? I’m happy to troubleshoot this issue with you in the sourcegraph slack (I can join the channel on Monday)....

One situation where we may need `SymbolInformation.owner` is to handle static inner java classes. The proposed solution for #1491 is to give the symbol syntax `java.util.Map.Entry#` for the inner class...

I have experienced the same problem, I am in favor of adding `.owner` but only fill it for local symbols 👍

Encountering the lack of `owner` when implementing rename in Metals for anonymous classes. A workaround is to filter out class symbols that contain the local method symbol as part of...

Thanks for reporting and providing the great reproduction! Something is going on with extension methods and I suspect it's related to https://github.com/scalameta/scalameta/issues/1780 With `logins()` ``` $ metap .bloop/finanzio/scala-2.12/classes/META-INF/semanticdb/finanzio/src/main/scala/finanzio/SaltedgeService.scala.semanticdb | grep...

Here's a minimized test case ```scala targeted( """package a |import scala.language.implicitConversions |class Ops[A](target: A) { | def map(fn: A => A): A = fn(target) |} |object ops { | implicit...

This is exciting! Minor thoughts: I like the CaseType encoding over using maps. TypeLambda should probably be called LambdaType for consistency with other names that also end with “Type”. Do...

I can live with a scheme where hardlinks are not globally unique as I agree it's not clear how to support that everywhere.

I don't expect regular users to encounter this problem since you need to drill into the internal `s.Type` API to experience this issue. From my limited experience so far squinting...