Am I not using higher-ordered-functions properly o...

# announcements

[Viktor Qvarfordt]

Am I not using higher-ordered-functions properly or is Kotlin’s type inference limited? (This is a silly example where I’m trying to get a “flow” or “left compose” operator - I know this is most natural with dot chaining in Kotlin.)

[Joffrey]

I'm not sure how the associativity works in your case, but I would say the error would make sense if the right-most expression is evaluated first. Try putting () around to see if that's the actual issue 😉

[Viktor Qvarfordt]

Good thinking, but this was not the case. I really wonder how to get this working 🤔

[Joffrey]

Actually I think map { it < 2 } may still be evaluated independently, despite your () , so maybe my initial solution was not a solution, but the problem may still be this one

[Joffrey]

You have to specify the type parameters explicitly for map , or maybe giving a type to the lambda's param is sufficient: map { i: Int -> i + 2 } . I'll let the smart guys answer the why :)

[Viktor Qvarfordt]

In principle, my code should behave just like ordinary .map, hopefully someone with deeper understanding of type inference catches this :)

[Joffrey]

I believe type inference is a bit harder for top-level functions like yours compared to functions with receivers like the stdlib's map

[Kroppeb]

This works fine for me.

[Viktor Qvarfordt]

@Kroppeb Interesting! That is precisely the code I have. What version of Kotlin and IntelliJ are you running?

[Kroppeb]

I did an update yesterday right before that. So kotlin 1.3.60 and latest intelliJ. But I didn't think the type inference got improved. Does your code compile if you try. Cause if compiling works then it might have been an issue with the ide that got fixed.

[Viktor Qvarfordt]

Thanks, I’ll have a look.

[Kroppeb]

When I wanted to reuse the scratch file for a different problem I noticed that my ide is now complaining again

[Kroppeb]

1.3.61

[Viktor Qvarfordt]

🤔 ok so this cannot be relied upon it seems