This is a basic setup in our work project. A HTTP ...

# coroutines

[gregorbg]

This is a basic setup in our work project. A HTTP Connector has to use a Thread pool of fixed size to execute Coroutine requests. The Foo implementation can do something outside of the suspend world.

1. Is it good practice / clean style to have the abstract connector provide its own CoroutineScope?
2. How can I better execute the blocking GET request? Explicitly passing the coroutineContext parameter "feels wrong" given that the other extension function builders "magically" supply it to the coroutine being created.

Overly_simplified_HTTP_connector_with_Coroutines.kt