I have a feature proposal, want to hear some feedback before posting to YT: Let's say we have this interface:

fun <E : Entity, ID : Serializable> getRequired(id: ID): E

We can use it either like a)

getRequired<SomeEntity,Long>(1L)

or b)

val entity: SomeEntity = getRequired(1L),

It seems to me that case a) is unnecessarily verbose, I'd like to write

getRequired<SomeEntity>(1L)

, because

1L

already contains information about type of the second generic parameter, so compiler can infer it, like it does in case b).

There's something like that https://youtrack.jetbrains.com/issue/KT-17061

thanks, I'll comment on that

there is always the possibility of an ugly workaround where you specify a useless second parameter that's just there to help type inference

The workaround can be useful if you specify second parameter as Class<E>

getRequired(1L, SomeEntity::class.java

and inside check for instance