Added lazy mountable extensions #3187
Conversation
| this.value = value | ||
| } | ||
|
|
||
| fun get() = value |
There was a problem hiding this comment.
Should this not be protected from incorrect access during initialisation? It can only be safely accessed from TestScope? That actually depends on how the underlying Extension initialises it, right?
This would work for my use-case, but the API is rather unsafe.
In the PR with the original use-case here there is LazyMaterialized with suspend fun get() so it can only be accessed from suspend TestScope.() -> Unit. If it's access before it needs to be somehow in a suspend context and will get loaded lazily then. If it's accessed after clear/close is called (on afterSpec or similar) then it'll throw an IllegalStateException.
Should we have a similarly "safer" API? I.e. the example here Kafka also involves some heavy I/O, blocking code both on start and close.
There was a problem hiding this comment.
So, short-answer: yes this works.
Long-answer: should we try to make the API safer?
There was a problem hiding this comment.
I agree the API isn't great, but I don't know how we can limit it to test scope, or even if we care about that, as long as it supports suspension.
Like you might want to initialize in a before spec listener. That's a suspendable context.
Let me think more on this.
There was a problem hiding this comment.
Ok I have followed your guidelines and it is much clearer now.
A LazyMountableExtension returns a LazyMaterialized which has a SAM.
interface LazyMaterialized<MATERIALIZED> {
suspend fun get(): MATERIALIZED
}Once get is invoked, the implementation can initialize if required.
Here is the example from the tests:
object : LazyMaterialized<String> {
var state: String? = null
override suspend fun get(): String {
delay(1)
if (state == null) state = "ready"
return state ?: error("Must be initialized")
}
}
|
Thank you for looking into this @sksamuel 🙏 |
Would something as simple as this suffice for your needs @nomisRev