Resource cache fails to complete future exceptionally & throws unhandled coroutine exception

[rli]

aws-toolkit-jetbrains/jetbrains-core/src/software/aws/toolkits/jetbrains/core/AwsResourceCache.kt

Lines 313 to 334 in 5c87fd9

	private fun <T> getCachedResource(context: Context<T>) {
	ApplicationManager.getApplication().executeOnPooledThread {
	var currentValue: Entry<T>? = null
	try {
	@Suppress("UNCHECKED_CAST")
	val result = cache.compute(context.cacheKey) { _, value ->
	currentValue = value as Entry<T>?
	fetchIfNeeded(context, currentValue)
	} as Entry<T>
	coroutineScope.launch {
	try {
	context.future.complete(result.value.await())
	} catch (e: Throwable) {
	val deferred = currentValue
	if (context.useStale && deferred != null && deferred.value.isCompleted) {
	context.future.complete(deferred.value.getCompleted())
	} else {
	context.future.completeExceptionally(e)
	}
	}
	}

L320 in DefaultAwsResourceCache can return a cached entry where the deferred value has completed exceptionally. As a result, result.value.await() (L325) throws, is caught by the surrounding block, and is unhandled due to the deferred.value.getCompleted() call on (L329).

Reproducable by doing a cache.getResource() call that throws the first time(successfully bubbles out to callee code), but the second call will trigger the bug

[rli]

Released in 1.44