I'm scratching my head here a bit regarding extension functions. Can someone explain to me why this works

private fun <String> HashMap<String, String>.myExt(key: String, incomingValue: String) {
    this[key] = incomingValue
}

But not this

private fun <String> HashMap<String, String>.myExt(key: String, incomingValue: String) {
    this[key] = "something else"
}

The latter gives me (in IntelliJ):

Type mismatch.
Required: String#1 (type parameter of myExt)
Found: kotlin.String

Interestingly,

this[key] = "something else" as String

fixes the compiler error?

Exactly, was just going to write that. So you think it's a bug? I was assuming that I'm missing something

Ah no, wait! It should be

fun HashMap<String, String>.myExt(key: String, incomingValue: String) {
    this[key] = "something else"
}

Of course! Thank you very much 🙂 It was autogenerated but that makes sense.

You're welcome 🙂

Regarding

String

vs

kotlin.String

I don't know really

the

String

in the original function is just a name you have given to the generic parameter, it has nothing to do with

kotlin.String

. You could replace it with

Abcd

(or more often used

T

) and the function would work exactly same

Now that you mention it, it's very clear 😃 thank you! 😃

It surprises me that that is even allowed, the potential for confusion is profound 😀

how would you prevent it tho? Ban all possible classnames?

Yeah I thought about that too, would be kind of tricky tro draw a line there

this is a pretty funny issue 🙂 I dunno about banning classnames, but any generic parameter that's not a single character should catch your eye. (I don't love that convention but it's the convention...)

I saw code using type parameters longer than a letter, always adopted the Uppercase convention, to avoid confusion. Example:

Message<REQUEST, RESPONSE>

or in Assertj:

interface Assert<SELF extends Assert<SELF, ACTUAL>, ACTUAL> ...