Events with subclasses

[vpratfr]

Hi,

first, thanks for this library. It's great. I have a question, I don't know whether this is intentional or a bug.

My Event hierarchy:

class A {
  public static class A1 extends A {}
  public static class A2 extends A {}
}

My producer:

class Activity {
  @Produce 
  public A1 initialA1Event() { return new A1; }

  // Register in onCreate, unregister in onDestroy
}

My receiver:

class Fragment {

  // Never gets the initial A1 event from activity
  @Subscribe
  public void receiveAnyA(A event) {}

  // Register in onResume, unregister in onPause
}

With the above code, the subscriber does not receive an initial A event which would be produced by the Activity. If I change the subscribe handler to receive A1 specifically, it works.

I would have expected that my Fragment gets notified of the activity's A1 as an A event.

[JakeWharton]

Producers don't honor class hierarchy like posting does. What would happen
if you were producing both A1 and A2?
On Jul 27, 2014 2:55 AM, "Vincent Mimoun-Prat" [email protected]
wrote:

Hi,

first, thanks for this library. It's great. I have a question, I don't
know whether this is intentional or a bug.

My Event hierarchy:

class A {
public static class A1 extends A {}
public static class A2 extends A {}
}

My producer:

class Activity {
@produce
public A1 initialA1Event() { return new A1; }

// Register in onCreate, unregister in onDestroy
}

My receiver:

class Fragment {

// Never gets the initial A1 event from activity
@subscribe
public void receiveAnyA(A event) {}

// Register in onResume, unregister in onPause
}

With the above code, the subscriber does not receive an initial A event
which would be produced by the Activity. If I change the subscribe handler
to receive A1 specifically, it works.

I would have expected that my Fragment gets notified of the activity's A1
as an A event.

—
Reply to this email directly or view it on GitHub
#122.

[vpratfr]

What would happen if you were producing both A1 and A2?

I would have expected in that case that both A1 and A2 get delivered to the subcribers of A. In fact I would expect the producers to simply have the same behaviour as when using posting.

[JakeWharton]

That seems illogical and doesn't scale to many subclasses. If you want a
produced value for A you should produce it directly and reconcile the many
different subclasses into a single value. Otherwise it sounds like you'd
want a different mechanism for seeding initial state.
On Jul 27, 2014 3:07 PM, "Vincent Mimoun-Prat" [email protected]
wrote:

What would happen if you were producing both A1 and A2?

I would have expected in that case that both A1 and A2 get delivered to
the subcribers of A.

—
Reply to this email directly or view it on GitHub
#122 (comment).

[alexustinovsm]

For instance,

public class A {
  // This code will work
  @Subscribe public onSomeEvent(SomeEvent event) { /* my code here */ }
}

// This code won't work because HandlerFinder doesn't find subscribe
public class A1 extends A {}

// This code won't work because HandlerFinder doesn't find subscribe
public class A2 extends A {}

I think this is a bug. I don't want to create the same method at 10 different children of A. I think it's very strange.

[sferra]

@alexustinovsm While the issue you describe doesn't relate to producers, I must agree that it is buggy behavior. The workaround would currently be to explicitly implement onSomeEvent in both A1and A2 so the methods are found:

public class A1 extends A {
  @Subscribe public onSomeEvent(SomeEvent event) {
    super.onSomeEvent(event);
  }
}

public class A2 extends A {
  @Subscribe public onSomeEvent(SomeEvent event) {
    super.onSomeEvent(event);
  }
}

Maybe open a separate ticket for this issue?

[adrien-aubel]

I agree with @sferra, this is a behavior that is surprising. I think this should be at least mentioned in the documentation.