1.x: Remove 'Completable' prefix from nested interfaces, move its subscription to top-level. #4425
Conversation
Current coverage is 84.23% (diff: 68.04%)@@ 1.x #4425 diff @@
==========================================
Files 270 270
Lines 17518 17548 +30
Methods 0 0
Messages 0 0
Branches 2677 2677
==========================================
- Hits 14785 14781 -4
- Misses 1877 1909 +32
- Partials 856 858 +2
|
|
Looks okay. What about the |
|
Oh, sure. I'll add them in real quick. I wasn't sure on your stance on them, but I think it'll make migration easier. |
|
Yes, having deprecated markers is more forgiveable than lacking the entire type. I don't know how extensively |
|
Updated! |
|
|
||
| /** | ||
| * Convenience interface and callback used by the lift operator that given a child CompletableSubscriber, | ||
| * return a parent CompletableSubscriber that does any kind of lifecycle-related transformations. | ||
| */ | ||
| public interface CompletableOperator extends Func1<CompletableSubscriber, CompletableSubscriber> { | ||
| public interface Operator extends Func1<rx.CompletableSubscriber, rx.CompletableSubscriber> { |
There was a problem hiding this comment.
I'm not certain about this being a compatible change.
So does this still compile?
CompletableOperator op = new CompletableOperator() {
@Override
public Completable.CompletableSubscriber call(Completable.CompletableSubscriber cs) {
return cs;
}
}There was a problem hiding this comment.
Hmm yeah it's not possible to provide perfect migration using this strategy. The covariant return type works but not the argument. We could provide overloads of lift, compose, and subscribe which accepted the old types, adapted them to the new type, and were marked as deprecated. That should provide perfect compatibility.
There was a problem hiding this comment.
In this case, how about the old ones not extending the new ones and have wrappers inside those methods of the old types?
There was a problem hiding this comment.
Updated with overloads
…tion to top-level.
|
Compilation issue resolved. |
|
I know, still waiting for a second approval. |
|
Okay, let's have this. One can complain anytime later. |
This is only the
Completablepart of #4420. It does not makeCompletableSubscriberan abstract class nor createSingle.Operator.