I would have expected the inliner to just copy the...

# announcements

[Bernhard]

I would have expected the inliner to just copy the code as is and file in all the type parameters with the real type variable

[diesieben07]

Can you clarify what you mean? Type parameters are still erased, there is no Class<T> or Class<String> after compilation, only Class.
But if you do T::class with a reified T, then you will get the proper Class instance.

[diesieben07]

You can't do that with a normal type parameter

[Bernhard]

exactly, you can do T::class

[Bernhard]

for instance in Gson you can parse a list of things by using:

mapper.fromJson<ArrayList<AdditionalFunction>>(it, object : TypeToken<ArrayList<AdditionalFunction>>() {}.type)

[Bernhard]

what I tried to do was to generalize it like this:

mapper.fromJson<ArrayList<T>>(it, object : TypeToken<ArrayList<T>>() {}.type)

[Bernhard]

where T is reified

[Bernhard]

and call it like this:

func<AdditionalFunction>()

[diesieben07]

That should work™ 😄

[Bernhard]

turn out it deserializes to a linked hashmap instead of a List<AdditionalFunction>

[Bernhard]

when you println what TypeToken<ArrayList<T>>() {}.type evaluates to, you get “ArrayList<T>>”

[Bernhard]

instead of “ArrayList<AdditionalFunction>”

[Bernhard]

so it feels like T is only reified if you access it via ::class

[Bernhard]

and inlining happens after type checking and is type erased

[diesieben07]

It seems like the compositing of types causes the issue here. The following prints `ArrayList<String>`:

private inline fun <reified R> a() {
    println(object : TypeToken<R>() {})
}

fun main(args: Array<String>) {
    a<ArrayList<String>>()
}

[Bernhard]

compiler bug?

[Bernhard]

or works as expected?

[diesieben07]

Not sure. Sounds like a bug to me, but I am not sure if it's possible to do properly.