I’ve got a sort of odd case here where I have a fu...

# eap

[Zac Sweers]

I’ve got a sort of odd case here where I have a function that returns Single<Collection<Type>> and internally code that looks like this

return someListSingle.map { list -> list as Collection<Type> }

In 1.8.0-RC, this cast is marked as redundant (as a compilation warning), but removing it actually causes a compilation error. I’m not sure if it’s expected to automatically infer the cast or if I should report a bug for a false-positive warning.

Screen Recording 2022-12-15 at 5.49.21 PM.mov

[Zac Sweers]

I see it with upcasting of any sort too, not just collections/generic types

[Ivan Kubyshkin [JetBrains]]

Can you attach a link to the code? Looks similar to https://youtrack.jetbrains.com/issue/KT-54027

[Zac Sweers]

looks the same but this passes with 1.7.22. Let me try to put together a simpler repro

[Zac Sweers]

Filed an issue with repros for all three issues (the above + two I described in this thread accidentally)

CC @mglukhikh @udalov

https://youtrack.jetbrains.com/issue/KT-55535/Type-inference-regressions-in-1.8.0-RC

[mglukhikh]

🙏