I'm developing a Java library that runs a scheduled task in a ScheduledExecutorService. If the user specifies callback(s), then after that task, those callback(s) will be invoked.
Right now I just have the scheduler thread invoking the callbacks. Should those callbacks execute in their own ExecutorService in their own thread pool?
In my opinion, the best approach would be to provide the client with both options: with an Executor specified for the callback and without it, similar to how it's done in CompletableStage:
public CompletionStage<Void> thenAcceptAsync(Consumer<? super T> action); // executes action on internal default executor
public CompletionStage<Void> thenAcceptAsync(Consumer<? super T> action, Executor executor); // executes action on provided executor
This way, the client can choose which approach is more convenient for their use case.
