@Scheduled not found when in parent classes

[edeandrea]

Describe the bug

When using the @Scheduled annotation on a method, it is not found/picked up when the scheduled method is on a parent class of a CDI bean.

Expected behavior

I would expect any @Scheduled method in the class hierarchy of a CDI bean to be found and executed.

Actual behavior

Only @Scheduled methods within the top-level class of a CDI bean are found/executed

How to Reproduce?

1. Download and unzip the reproducer: scheduled.zip
2. Examine the project.
   • There is a class called BeanWithScheduled which is simple CDI bean with an @Scheduled method.
   • There is a class called ChildBean which has an @Scheduled method and also extends a base class, SubclassWithScheduled.
   • SubclassWithScheduled also has an @Scheduled method.
3. Run ./mvnw clean quarkus:dev
4. You'll notice the following output, indicating that only the @Scheduled methods in the BeanWithScheduled and ChildBean classes were executed. The @Scheduled method in the SubclassWithScheduled base class was not executed.

__  ____  __  _____   ___  __ ____  ______ 
 --/ __ \/ / / / _ | / _ \/ //_/ / / / __/ 
 -/ /_/ / /_/ / __ |/ , _/ ,< / /_/ /\ \   
--\___\_\____/_/ |_/_/|_/_/|_|\____/___/   
2024-02-14 09:39:21,697 INFO  [io.quarkus] (Quarkus Main Thread) scheduled 1.0-SNAPSHOT on JVM (powered by Quarkus 3.7.2) started in 0.563s. 

2024-02-14 09:39:21,698 INFO  [io.quarkus] (Quarkus Main Thread) Profile dev activated. Live Coding activated.
2024-02-14 09:39:21,698 INFO  [io.quarkus] (Quarkus Main Thread) Installed features: [cdi, scheduler, smallrye-context-propagation, vertx]
2024-02-14 09:39:22,022 INFO  [BeanWithScheduled] (vert.x-worker-thread-1) Running scheduled task from main bean without a parent class
2024-02-14 09:39:22,022 INFO  [ChildBean] (vert.x-worker-thread-2) Running scheduled task from main bean with a parent class
2024-02-14 09:39:27,010 INFO  [ChildBean] (vert.x-worker-thread-2) Running scheduled task from main bean with a parent class
2024-02-14 09:39:27,010 INFO  [BeanWithScheduled] (vert.x-worker-thread-1) Running scheduled task from main bean without a parent class

Output of uname -a or ver

Darwin edeandrea-m1pro 23.3.0 Darwin Kernel Version 23.3.0: Wed Dec 20 21:30:44 PST 2023; root:xnu-10002.81.5~7/RELEASE_ARM64_T6000 arm64

Output of java -version

openjdk version "17.0.10" 2024-01-16
OpenJDK Runtime Environment Temurin-17.0.10+7 (build 17.0.10+7)
OpenJDK 64-Bit Server VM Temurin-17.0.10+7 (build 17.0.10+7, mixed mode)

Quarkus version or git rev

3.7.2

Build tool (ie. output of mvnw --version or gradlew --version)

Apache Maven 3.9.6 (bc0240f3c744dd6b6ec2920b3cd08dcc295161ae)
Maven home: /Users/edeandre/.m2/wrapper/dists/apache-maven-3.9.6-bin/3311e1d4/apache-maven-3.9.6
Java version: 17.0.10, vendor: Eclipse Adoptium, runtime: /Users/edeandre/.sdkman/candidates/java/17.0.10-tem
Default locale: en_US, platform encoding: UTF-8
OS name: "mac os x", version: "14.3", arch: "aarch64", family: "mac"

Additional information

No response

[quarkus-bot]

/cc @mkouba (scheduler)

[mkouba]

Well, I don't remember all the details but we initially decided not to support the inheritance of @Scheduled methods.

It would definitely make the discovery and validation more complex. Also if there are multiple subclasses inheriting a scheduled method from a superclass you wouldn't be able to declare the @Scheduled#identity() which has to be unique.

As a workaround, you can override the scheduled method in the subclass (ChildBean in your reproducer) and re-declare the @Scheduled annotation 🤷.

[edeandrea]

Yeah I had just thought of that workaround and thats most likely what I'll do. I don't really know the history as to why it was excluded, but I'm sure there are pros/cons on each side of including it or not. I was just surprised that it didn't work and spent half a day trying to figure out why my code wasn't working.

If it isn't going to be supported in parent classes, maybe that should be called out in the docs?

[mkouba]

If it isn't going to be supported in parent classes, maybe that should be called out in the docs?

That's a good idea. Actually, we can also detect the case where a @Scheduled method is declared on an abstract class (your reproducer)/interface and fail the build. I'll send a PR.

[mkouba]

Hm, actually it is documented in the reference guide - there is a warning that starts with text "Subclasses never inherit the metadata of a @Scheduled method declared on a superclass.". It got added in #24270 which a fix for #24212 and removes the inheritance for the reasons described in the comments.

So I will:

1. Turn the warning into a subsection with title Inheritance of metadata,
2. Mention this in the javadoc of io.quarkus.scheduler.Scheduled as well,
3. Implement a validation rule to fail the build if @Scheduled is declared on a method of an abstract class/interface

[edeandrea]

Is it just when declared on a method of an abstract class? Or anywhere in the parent class hierarchy?

[mkouba]

Is it just when declared on a method of an abstract class? Or anywhere in the parent class hierarchy?

An abstract class or an interface. A superclass that is not abstract might be a legal use case.

[edeandrea]

Thanks very much @mkouba for your help!