It would be nice to have a context-receiver "summo...

# stdlib

[Youssef Shoaib [MOD]]

It would be nice to have a context-receiver "summoning" function in the stdlib:

context(A) fun <A> given(): A = this@A

to be used like given<List<String>>(). I think this is better than the this@ syntax since it supports generics (like List<String>) and can "summon" supertypes of a context receiver which gets resolved to the apt subtype context

Edit: I realise that maybe stdlib can't have context declarations. Instead, it can be defined equivalently as:

fun <A> A.given(): A = this

The only issue is that it pollutes the namespace of Any.

[ephemient]

I haven't thought about it much but would https://github.com/Kotlin/KEEP/blob/statics/proposals/statics.md be workable?

context(T)
val T.static.context: T
    get() = this@T

List<String>.context

[ephemient]

that would prevent things like

f(given(), given())

from inferring types, which I would find kind of sketchy with context receivers

[Javier]

Without knowing if that works, I think given<A>() is more idiomatic.

But the issue you mentioned can have sense

[Youssef Shoaib [MOD]]

f(given(), given()) doesn't actually infer types. An unspecified given() always resolves to the closest receiver. If the closest receiver is part of a receiver group (i.e. the closest receiver is context(A, B)), then you get an error. Personally, I would've liked if f(given(), given()) would've worked, but I see the ambiguity and resolution issues with it.