search

vsilaev / tascalate-async-await

Async / Await asynchronous programming model for Java versions 1.8 - 23; similar to the functionality available in C# 5. The implementation is based on continuations for Java (see my other projects).
BSD 2-Clause "Simplified" License
104 stars 12 forks source link
async async-await asynchronous-programming concurrent-programming continuations coroutines java

readme

Maven Central GitHub release license

Why async-await?

Asynchronous programming has long been a useful way to perform operations that don’t necessarily need to hold up the flow or responsiveness of an application. Generally, these are either compute-bound operations or I/O bound operations. Compute-bound operations are those where computations can be done on a separate thread, leaving the main thread to continue its own processing, while I/O bound operations involve work that takes place externally and may not need to block a thread while such work takes place. Common examples of I/O bound operations are file and network operations.

Traditional asynchronous programming involves using callbacks these are executed on operation completion. The API differs across languages and libraries, but the idea is always the same: you are firing some asynchronous operation, get back some kind of Promise as a result and attach success/failure calbacks that are executed once asynchronous operation is completed. However, these approach is associated numerous hardships:

  1. You should explicitly pass contextual variables to callbacks. Sure, you can use lambdas to capture lexical context, but this does not eliminate the problem completly. And sometimes even sacrifies readability of the code - when you have a lot of lambda functions with complex body.
  2. Coordination of asynchronous operations with callbacks is dificult: any branching logic inside the chain of asynchronous callbacks is a pain; resource management provided by try-with-resources constructs are not possible with asynchronous callbacks as well as many other control flow statements; handling failures is radically different from the familiar try/catch used in synchronous code.
  3. Different callbacks are executed on different threads. Hence special care should be taken where the application flow resumes. The issue is very critical when application runs in managed environment like JEE or UI framework (JavaFX, Swing, etc).

To alleviate aforementioned readability and maintainability issues some languages provides async/await asynchronous programming model. This lets developer make asynchronous calls just as easily as she can invoke synchronous ones, with the tiny addition of a keyword await and without sacrifying any of asynchronous programming benefits. With await keyword asynchronous calls may be used inside regular control flow statements (including exception handling) as naturally as calls to synchronous methods. The list of the languages that support this model is steadly growing: C# 5, ECMAScript 7, Kotlin, Scala.

Tascalate Async/Await library enables async/await model for projects built with the Java 8 and beyond. The implementation is based on continuations for Java and provides runtime API + bytecode enchancement tools to let developers use syntax constructs similar to C# 5 or ECMAScript 2017/2018 with pure Java.

How to use ?

...with Maven

First, add Maven dependency to the library runtime:

<dependency>
    <groupId>net.tascalate.async</groupId>
    <artifactId>net.tascalate.async.runtime</artifactId>
    <version>1.2.7</version>
</dependency>

Second, add the following build plugins in the specified order:

<build>
  <plugins>

    <plugin>
      <groupId>net.tascalate.async</groupId>
      <artifactId>net.tascalate.async.tools.maven</artifactId>
      <version>1.2.7</version>
      <executions>
        <execution>
          <id>tascalate-async-enhance-main-classes</id> 
          <phase>process-classes</phase>
          <goals>
            <goal>tascalate-async-enhance</goal>
          </goals>
        </execution>
    <!-- Only if you need to enhance test classes -->      
        <execution>
          <id>tascalate-async-enhance-test-classes</id> 
          <phase>process-test-classes</phase>
          <goals>
            <goal>tascalate-async-enhance</goal>
          </goals>
        </execution>
      </executions>
    </plugin>

    <plugin>
      <groupId>net.tascalate.javaflow</groupId>
      <artifactId>net.tascalate.javaflow.tools.maven</artifactId>
      <version>2.7.6</version>
      <executions>
        <execution>
          <id>javaflow-enhance-main-classes</id> 
          <phase>process-classes</phase>
          <goals>
            <goal>javaflow-enhance</goal>
          </goals>
        </execution>
        <!-- Only if you need to enhance test classes -->       
        <execution>
          <id>javaflow-enhance-test-classes</id> 
          <phase>process-test-classes</phase>
          <goals>
            <goal>javaflow-enhance</goal>
          </goals>
        </execution>
      </executions>
    </plugin>

  </plugins>
</build>

You are ready to start coding!

...with Gradle

As with Maven, you have to specify both build plugins and runtime dependencies. The minimal Gradle scipt should have the following prologue:

buildscript {
    repositories {
        mavenCentral()
    }

    dependencies {
        classpath 'net.tascalate.async:net.tascalate.async.tools.gradle:1.2.7'
        classpath 'net.tascalate.javaflow:net.tascalate.javaflow.tools.gradle:2.7.6'
        /* other plugins */
    }
}

apply plugin: "java"
/* ORDER IS IMPORTANT: Async/Await before Continuations! */
apply plugin: "async-await"
apply plugin: "continuations"

repositories {
    mavenCentral()
}

dependencies {
    implementation 'net.tascalate.async:net.tascalate.async.runtime:1.2.7'
    /* other dependencies */
}

The more advanced example with Async/Await Extras module + Tascalate Concurrent and Async/Await SchedulerResolver-s (discussed below) will be:

buildscript {
    repositories {
        mavenCentral()
    }

    dependencies {
        classpath 'net.tascalate.async:net.tascalate.async.tools.gradle:1.2.7'
        classpath 'net.tascalate.javaflow:net.tascalate.javaflow.tools.gradle:2.7.6'
        /* other plugins */
    }
}

apply plugin: "java"
/* ORDER IS IMPORTANT: Async/Await before Continuations! */
apply plugin: "async-await"
apply plugin: "continuations"

repositories {
    mavenCentral()
}

dependencies {
    implementation 'net.tascalate.async:net.tascalate.async.runtime:1.2.7'

    /* Async/Await Extras */
    implementation 'net.tascalate.async:net.tascalate.async.extras:1.2.7'

    /* Promise<T> implementation */
    /* Necessary because net.tascalate.async.extras uses it as an */
    /* 'optional' dependency to avoid concrete version lock-in.   */
    implementation 'net.tascalate:net.tascalate.concurrent:0.9.8'

    /* Necessary only for different providers */
    runtimeOnly 'net.tascalate.async:net.tascalate.async.resolver.provided:1.2.7'
    /*
    runtimeOnly 'net.tascalate.async:net.tascalate.async.resolver.propagated:1.2.7'
    */

    /* other dependencies */
}
/* Optional config */
'async-await' {
    /* ... */
}

'continuations' {
    /* ... */
}

Asynchronous tasks

The first type of functions the library supports is asycnhronous task. Asynchronous task is a method (either instance or class method) that is annotated with net.tascalate.async.async annotation and returns CompletionStage<T> or void. In the later case it is a "fire-and-forget" task that is intended primarly to be used for event handlers inside UI framework (like JavaFX or Swing). Let us write a simple example:

import static net.tascalate.async.CallСontext.async;
import static net.tascalate.async.CallСontext.await;
import net.tascalate.async.async;

import java.util.concurrent.CompletionStage;
import java.util.concurrent.CompletableFuture;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;

class MyClass {
    public @async CompletionStage<String> mergeStrings() {
        StringBuilder result = new StringBuilder();
        for (int i = 1; i <= 10; i++) {
            String v = await( decorateStrings(i, "async ", " awaited") );
            result.append(v).append('\n');
        }
        return async(result.toString());
    }

    public @async CompletionStage<String> decorateStrings(int i, String prefix, String suffix) {
        String value = prefix + await( produceString("value " + i) ) + suffix;
        return async(value);
    }

    // Emulate some asynchronous business service call
    private static CompletionStage<String> produceString(String value) {
        return CompletableFuture.supplyAsync(() -> value, executor);
    }

    private static final ExecutorService executor = Executors.newFixedThreadPool(4);
}

Thanks to statically imported methods of net.tascalate.async.CallСontext the code looks very close to the one developed with languages having native support for async/await. Both mergeStrings and decorateStrings are asynchronous methods -- they are marked with net.tascalate.async.async annotation and returns CompletionStage<T>. Inside these methods you may call await to suspend the method till the CompletionStage<T> supplied as the argument is resolved (either sucessfully or exceptionally). Please notice, that you can await for any CompletionStage<T> implementation obtained from different libraries - like inside the decorateStrings method, including pending result of another asynchronous method - like in mergeStrings.

The list of the supported return types for the async methods is:

  1. void
  2. java.util.concurrent.CompletionStage
  3. java.util.concurrent.CompletableFuture
  4. net.tascalate.concurrent.Promise (see my other project Tascalate Concurrent)

For non-void results the actual result type class also implements java.util.concurrent.Future (even for the case [2] with CompletionStage). This means that you can safely upcast the result promise to the java.util.concurrent.Future and use blocking methods if necessary. Most importantly, you can use the cancel(...) method cancel the future returned.

To return a result from the asynchronous method you have to use syntactic construct return async(value). You must always treat both of these statements (calling async method and return-ing its result) as the single syntactic construct and don't call async method separately or store it return value to variable while these will lead to unpredicatble results. It's especially important if your method body is not linear. Depending on your established coding practice how to deal with multiple returns you should use either...

public @async CompletionStage<String> foo(int i) {
    switch (i) {
        case 1: return async("A");
        case 2: return async("B");
        case 3: return async("C");
        default:
            return async("<UNKNOWN>");
    }
}

...or...

public @async CompletionStage<String> bar(int i) {
    String result;
    switch (i) {
        case 1: result = "A"; break;
        case 2: result = "B"; break;
        case 3: result = "C"; break;
        default:
            result = "<UNKNOWN>";
    }
    return async(result);
}

It's worth to mention, that when developing code with async/await you should avoid so-called "async/await hell". In short, pay special attention what parts of your code may be executed in parallel and what parts require serial execution. Consider the following example:

public @async CompletionStage<Long> calculateTotalPrice(Order order) {
    Long rawItemsPrice = await( calculateRawItemsPrice(order) );  
    Long shippingCost  = await( calculateShippingCost(order) );  
    Long taxes         = await( calculateTaxes(order) );  
    return async(rawItemsPrice + shippingCost + taxes);
}

protected @async CompletionStage<Long> calculateRawItemsPrice(Order order) {
    ...
}

protected @async CompletionStage<Long> calculateShippingCost(Order order) {
    ...
}

protected @async CompletionStage<Long> calculateTaxes(Order order) {
    ...
}

In the above example all async methods calculateRawItemsPrice, calculateShippingCost, calculateTaxes are executed serially, one by one, hence the performance is degraded comparing to the following parallelized solution:

public @async CompletionStage<Long> calculateTotalPrice(Order order) {
    CompletionStage<Long> rawItemsPrice = calculateRawItemsPrice(order);  
    CompletionStage<Long> shippingCost  = calculateShippingCost(order);  
    CompletionStage<Long> taxes         = calculateTaxes(order);  
    return async( await(rawItemsPrice) + await(shippingCost) + await(taxes) );
}

This way all inner async operations are started (almost) simualtenously and are running in parallel, unlike in the first example.

Suspendable methods

Sometimes it is necessary to await for asynchronous result in some helper method that per se should not be asynchronous. To support this use case Tascalate Async/Await provides @suspendable annotation. The original example above hence can be rewritten as following:

import static net.tascalate.async.CallСontext.async;
import static net.tascalate.async.CallСontext.await;
import net.tascalate.async.async;
import net.tascalate.async.suspendable; // NEW ANNOTATION IMPORT

import java.util.concurrent.CompletionStage;
import java.util.concurrent.CompletableFuture;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;

class MyClass {
    public @async CompletionStage<String> mergeStrings() {
        StringBuilder result = new StringBuilder();
        for (int i = 1; i <= 10; i++) {
        // No await here -- moved to helper method
            String v = decorateStrings(i, "async ", " awaited"); 
            result.append(v).append('\n');
        }
        return async(result.toString());
    }

    // Async method refactored to suspendable
    private @suspendable String decorateStrings(int i, String prefix, String suffix) {
        String value = prefix + await( produceString("value " + i) ) + suffix;
        return value; // Just regular "return <value>" instead of "return async(<value>)"
    }

    // Emulate some asynchronous business service call
    private static CompletionStage<String> produceString(String value) {
        return CompletableFuture.supplyAsync(() -> value, executor);
    }

    private static final ExecutorService executor = Executors.newFixedThreadPool(4);
}

As you see, suspendable methods are just like regular ones but with special annotation - @suspendable. You should follow regular rules about returning results from this methods, moreover - it's an error to call return async(<value>) inside these methods. The important thing about @suspendable methods is that they may be called only from @async methods or from other @suspendable methods.

Performance-wise suspendable methods behaves the same as asynchronous task methods, so the question "which kind should be used" is justy a matter of orginizing and structuring your code . The recommended approach is to use asynchronous task methods when they are exposed to outside clients and suspendable ones for internal implementation details. However, the final decision is up to library user till s/he holds the rule that suspendable methods may be called only from asynchronous context (@async methods or other @suspendable methods) as stated above.

Implemenation notes: technically suspendable methods are implemented as continuable methods that follow rules defined by Tascalate JavaFlow library, so you may use any continuable annotation that is supported by Tascalate JavaFlow, not only @suspendable.

Generators

TDB

Scheduler & SchedulerResolver - where is my code executed?

Introducing schedulers

When executing asynchronous code with CompletionStage / CompletableFuture it's critical to know where the code is resumed once the corresponding completion stage is settled. With regular CompletionStage API the answer is pretty straightforward: the code will be resumed with the Executor supplied as an additional parameter to the API method like below:

import java.util.concurrent.CompletionStage;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
...

CompletionStage<String> myCompletionStage = ... ; // Start asynchronous operation
ExecutorService myExecutor = Executors.newFixedThreadPool(4);
myCompletionStage.thenAcceptAsync(System.out::println, myExecutor);

In the example above the code to print result to the console will run on the thread provided by myExecutor.

However, for Tascalate Async / Await there is no way to specify explicitly where the code will be resumed once the corresponding await(future) operation is complete. Instead of the passing Executor explicitly, the library uses more declarative pluggable mechanism to specify asynchronous executor to run with.

First we must introduce the Scheduler interface:

package net.tascalate.async;

public interface Scheduler {

   ...    

    default Runnable contextualize(Runnable resumeContinuation) {
        return resumeContinuation;
    }

    abstract public CompletionStage<?> schedule(Runnable runnable);
   ...
}

The Schedulre API has 2 responsibilities:

  1. Execute supplied runnable command, most probably asynchronously (thought, this is implementation-dependent)
  2. Capture execution context of the currently running thread before suspension so it can later be restored when the code is resumed after await(future). The execution context is typically defined as a set of thread local variables - either via explicit usage of the ThreadLocal or via some API wrapping ThreadLocal-s (like SimpAttributesContextHolder in Spring)

There are several factory methods in Scheduler interface that create concrete Scheduler implementation using the ExecutorService supplied and optional contextualizer - a function that captures current thread execution context and creates a wrapper for the runnable to re-apply context on the new thread.

package net.tascalate.async;
...
import java.util.concurrent.Executor;
import java.util.concurrent.ExecutorService;
import java.util.function.Function;
...

public interface Scheduler {
    ...

    public static Scheduler nonInterruptible(Executor executor);

    public static Scheduler nonInterruptible(
        Executor executor, 
        Function<? super Runnable, ? extends Runnable> contextualizer);     

    public static Scheduler interruptible(ExecutorService executor);

    public static Scheduler interruptible(
        ExecutorService executor, 
        Function<? super Runnable, ? extends Runnable> contextualizer);

   ...
}

So the sequence to create Scheduler is:

import java.util.concurrent.CompletionStage;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;

import java.util.function.Function;

import net.tascalate.async.Scheduler 
...
ExecutorService myExecutor = Executors.newFixedThreadPool(4); // Or whatever ExecutorService impl. you need
Scheduler myScheduler = Scheduler.interruptible(myExecutor, Function.identity());

The careful reader may notice that there is a divide between interruptible vs non-interruptible schedulers, but let us left this out of the scope for a while. Instead, let's discuss how to apply the scheduler created to the asynchronous methods.

Explicit (method-scoped) schedulers

The most explicit and straightforward way to specify asynchronous Scheduler is to pass it explicitly to the asynchronous method as an annotated parameter:

import static net.tascalate.async.CallСontext.async;
import static net.tascalate.async.CallСontext.await;
import net.tascalate.async.async;
import net.tascalate.async.Scheduler;
import net.tascalate.async.SchedulerProvider; 

import java.util.concurrent.CompletionStage;
import java.util.concurrent.CompletableFuture;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;

public class MyClass {
   public static void main(String[] argv) {
       ExecutorService myExecutor = Executors.newFixedThreadPool(4);
       Scheduler myScheduler = Scheduler.interruptible(myExecutor);
       CompletionStage<String> myPromise = new MyClass().mergeStrings(myScheduler);
   }

    public @async CompletionStage<String> mergeStrings(@SchedulerProvider currentScheduler) {
        StringBuilder result = new StringBuilder();
        for (int i = 1; i <= 10; i++) {
            String v = await( decorateStrings(i, "async ", " awaited", currentScheduler) );
            result.append(v).append('\n');
        }
        return async(result.toString());
    }

    public @async CompletionStage<String> decorateStrings(int i, String prefix, String suffix, 
        @SchedulerProvider currentScheduler) {

        String value = prefix + await( produceString("value " + i) ) + suffix;
        return async(value);
    }

    // Emulate some asynchronous business service call
    private static CompletionStage<String> produceString(String value) {
        ...
    }
}

The scenario is simple: add parameter of the type Scheduler annotated with @SchedulerProvider annotation to the each asynchronous method where this scheduler is used and pass it explicitly. Note that it's an error to have more than one parameter annotated with @SchedulerProvider - only one is allowed. Also, passing non-annotated Scheduler will have no effect -- it's treated just like regular parameter and do not used for asynchronous execution.

You may notice that currentScheduler parameter from the mergeStrings method is passed directly to the decorateStrings method. This is mandatory if you want to share the same scheduler accross several asynchronous methods. By default schedulers are not inherited for nested calls.

Notice that in both methods above currentScheduler is not used with await(...) operator - it's used implicitly behinds the scenes in the generated code. This has one important implication: you can use only one scheduler per asynchronous method, there is no way to use different schedulers for different await(...) operations within the same method. If you ever need to then please re-factor your code to use separate asynchronous methods where individual schedulers may be defined per-method.

SchedulerResolver introduction -- propagating scheduler to nested calls

As it was mentioned right above "by default schedulers are not inherited for nested calls". This is a good point to introduce pluggable scheduler providers mechanism and alleviate this limitation.

To use pluggable schedule provider you need to add corresponding Maven dependency and introduce new artifact on the project class-path / module-path. For example:

<dependency>
    <groupId>net.tascalate.async</groupId>
    <artifactId>net.tascalate.async.resolver.propagated</artifactId>
    <version>${actual-tascalate-async-await-version}</version>
    <scope>runtime</scope>
</dependency>

Now let us rewrite the example above to automatically propagate scheduler of the outer asynchronous method to the inner one:

import static net.tascalate.async.CallСontext.async;
import static net.tascalate.async.CallСontext.await;
import net.tascalate.async.async;
import net.tascalate.async.Scheduler;
import net.tascalate.async.SchedulerProvider; 
import net.tascalate.async.spi.CurrentCallContext;

import java.util.concurrent.CompletionStage;
import java.util.concurrent.CompletableFuture;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;

public class MyClass {
   public static void main(String[] argv) {
       ExecutorService myExecutor = Executors.newFixedThreadPool(4);
       Scheduler myScheduler = Scheduler.interruptible(myExecutor);
       CompletionStage<String> myPromise = new MyClass().mergeStrings(myScheduler);
   }

    public @async CompletionStage<String> mergeStrings(@SchedulerProvider Scheduler currentScheduler) {
        System.out.println("Current scheduler (outer) - " + CurrentCallContext.scheduler());
        StringBuilder result = new StringBuilder();
        for (int i = 1; i <= 10; i++) {
            String v = await( decorateStrings(i, "async ", " awaited") );
            result.append(v).append('\n');
        }
        return async(result.toString());
    }

    public @async CompletionStage<String> decorateStrings(int i, String prefix, String suffix) {
        System.out.println("Current scheduler (inner) - " + CurrentCallContext.scheduler());
        String value = prefix + await( produceString("value " + i) ) + suffix;
        return async(value);
    }

    // Emulate some asynchronous business service call
    private static CompletionStage<String> produceString(String value) {
        ...
    }
}

You may see that the code was simplified, however no new specific code for the "propagating provider" was introduced. If you run this code you will see that CurrentCallContext.scheduler() reports the very same scheduler for both inner and outer methods. Btw, CurrentCallContext.scheduler() may be used with any combination of the scheduler providers and reports currently used Scheduler for all asynchronous, suspendable and generators methods.

More useful SchedulerResolver - per-class/per-instance schedulers

The next and probably the most useful one Scheduler provider is a "provided" provider variant (no pun). The idea is that a Scheduler may be specified per class or per class instances as a filed (an instance one or a static one) or as a getter-like method (no argument method with a Scheduler return type).

To use this provider you first need to add a new runtime dependency:

<dependency>
    <groupId>net.tascalate.async</groupId>
    <artifactId>net.tascalate.async.resolver.provided</artifactId>
    <version>${actual-tascalate-async-await-version}</version>
    <scope>runtime</scope>
</dependency>

Let us modify an example from the above to use the new provider:

import static net.tascalate.async.CallСontext.async;
import static net.tascalate.async.CallСontext.await;
import net.tascalate.async.async;
import net.tascalate.async.Scheduler;
import net.tascalate.async.SchedulerProvider; 
import net.tascalate.async.spi.CurrentCallContext;

import java.util.concurrent.CompletionStage;
import java.util.concurrent.CompletableFuture;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;

public class MyClass {
   public static void main(String[] argv) {
       ExecutorService myExecutor = Executors.newFixedThreadPool(4);
       Scheduler myScheduler = Scheduler.interruptible(myExecutor);
       CompletionStage<String> myPromise = new MyClass(myScheduler).mergeStrings();
   }

   @SchedulerProvider /* Mandatory annotation */
   private final Scheduler scheduler;

   MyClass(Scheduler scheduler) {
     this.scheduler = scheduler;
   }

    public @async CompletionStage<String> mergeStrings( ) {
        System.out.println("Current scheduler (outer) - " + CurrentCallContext.scheduler());
        StringBuilder result = new StringBuilder();
        for (int i = 1; i <= 10; i++) {
            String v = await( decorateStrings(i, "async ", " awaited") );
            result.append(v).append('\n');
        }
        return async(result.toString());
    }

    public @async CompletionStage<String> decorateStrings(int i, String prefix, String suffix) {
        System.out.println("Current scheduler (inner) - " + CurrentCallContext.scheduler());
        String value = prefix + await( produceString("value " + i) ) + suffix;
        return async(value);
    }

    // Emulate some asynchronous business service call
    private static CompletionStage<String> produceString(String value) {
        ...
    }
}

The Scheduler is provided as an instance field for all of @async instance methods of the class MyClass. You can initialize this variable in the constructor (as above) or at any time before invoking the @async method. In Spring / CDI environment the scheduler field might be injected by the container via corresponding annotation (@Autowired or @Injected).

Please notice that when you are re-assigning the field during execution of the @async method it has no effect on the methods these are in progress -- only freshly invoked ones will see the change. However, special consideration should be taken on account: in the example above if you re-define the scheduler field after mergeStrings invocation but before decorateStrings invocation then methods will use different schedulers. Also, no special synchronization is performed by the library itself, and it's library's user responsibility to synchronize access to such fields.

As it was mentioned, you can use a getter-like method annotated with @SchedulerProvider to supply scheduler. Use this option when you need different schedulers for the different object states, but, again, provide all necessary state synchronization on your own.

It's an error to provide a Scheduler with both a field and a method, or to have more than one filed or more than one getter-like method annotated with @SchedulerProvider.

It was mentioned that you can use both instance and static (class) field / method to provide a Scheduler. However, consider the following rules:

  1. Instance-level provider supplies a Scheduler only to @async instance methods.
  2. Class-level provider supplies a Scheduler to static @async methods AND to instance methods UNLESS there is a separate instance-level provider.
  3. It's an error to have more than one class-level provider in the same class via static field(s) / static getter-like method(s) / the combination of thereof (same as with instance-level providers); however it's a fully supported scenario when you have both instance-level provider AND class-level provider: instance level provider will take precedence over the class-level provider for the @async instance methods.

Last but not least is a visibility of the Scheduler provider (field / getter-like method) inherited from the superclass. It follows the same visibility rules as for the regular fields / methods inheritance: public and protected are always visible; package private are visible when both classes are in the same package; and private members are not visible. Take this on account when runtime will report you about ambiguity of the Scheduler provider - most probably, your subclass inherits ones from the superclasses chain.

Scoped SchedulerResolver -- overriding schedulers, providing own schedulers in DI environment

TBD

Interruptions/cancelation of @async methods & exception handling

TBD