IntelliJ (and dokka soon) shows a warning when using `suspend fun CoroutineScope.foo() { }`: `Ambigu...

# coroutines

[hfhbd]

IntelliJ (and dokka soon) shows a warning when using `suspend fun CoroutineScope.foo() { }`: Ambiguous coroutineContext due to CoroutineScope receiver of suspend function, because CoroutineScope.coroutineContext and coroutineContext. But why does coroutines-core itself use this construct? https://github.com/Kotlin/kotlinx.coroutines/blob/52cbf01c63c0737ee745ff08568f26385271a9ab/kotlinx-coroutines-core/common/src/Builders.common.kt#L50

[ephemient]

launch is a scope builder

[hfhbd]

Yeah, I mean the lambda parameter too

[ephemient]

it builds a new coroutine scope, which is both necessary within the lambda to perform more operations (launch etc.) and shadows and outer scope

[hfhbd]

But this also shadows coroutineContext from the suspend lambda and, according to the hint, you should use coroutineScope instead

[ephemient]

yes it would be a possible API design to not provide a CoroutineScope receiver to the lambda. but this would be more error-prone

[hfhbd]

Okay, so it is save to use a suspend CoroutineScope.() -> Unit lambda, but is not safe to use suspend fun CoroutineScope.foo() and does not result into shadowing coroutineContext?