How I currently use structured concurrency

[Josiah Noel]

Heya, not exactly sure if this is the place for this, but thought I'd share
how I'm currently using SC with JDK 20 as a backend dev.

For some non-critical services and some side projects, I've been refactored
from reactive to imperative style. Normally when I use SC it's to
orchestrate multiple API/DB calls and to cancel the tasks if a subtasks
fail. The underlying resources are usually managed by my DI framework
(usually Spring or Avaje), so I don't really have custom logic to clean up
resources.

The only (admittedly minor) complaints I have are the lack of a
runnable fork variation and the annoying checked exceptions. Other than
that, works like a dream.

Below is a ShutdownOnFailure wrapper class I typically use to get around
the 2 complaints.

public class UncheckedShutdownOnFailure implements AutoCloseable {


private final ShutdownOnFailure scope = new ShutdownOnFailure();


public <T> Future<T> fork(Callable<? extends T> task) {


return scope.fork(task);

}


public Future<Object> fork(Runnable task) {


return scope.fork(Executors.callable(task));

}


public void join() {


try {


scope.join().throwIfFailed(this::uncheck);


} catch (final InterruptedException e) {


Thread.currentThread().interrupt();


throw new IllegalStateException(e);

}

}

private RuntimeException uncheck(Throwable t) {


if (t instanceof final RuntimeException ex) {

return ex;

}


return new IllegalStateException(t);

}


@Override

public void close() {


scope.close();

}

}

-- 
Cheers, Josiah.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/loom-dev/attachments/20230514/35693ec2/attachment.htm>