stream-testkit: issue with scala 2.12 usage of expectNextWithTimeoutPF

[pjfanning]

see apache/pekko-persistence-cassandra#212
there change that seems to have caused the issue is #1218 (and earlier #1186)

[pjfanning]

@He-Pin what do you think we should do here? Should we document the issue we see in apache/pekko-persistence-cassandra#212 or is there a change we can make to stream-testkit to avoid the issue going forward?

[He-Pin]

I was expected to avoid a double evaluation, but that can only be done with 2.13, we may better go with 2.13 abd I expect drop the 2.12 support in 1.2.x anyway

[pjfanning]

@He-Pin This new method in https://github.com/He-Pin/incubator-pekko/blob/05248075bd13a336c9af079cb44a61986dc1f751/stream-testkit/src/main/scala/org/apache/pekko/stream/testkit/StreamTestKit.scala#L869 doesn't make sense to me. The PartialFunction is a Scala function, not a Java function. This seems like a not very useful Java API. The other changes in #1186 seem mainly ok to me but could we consider these?

• removing the Java API expectNextWithTimeoutPF
• or changing the Java API expectNextWithTimeoutPF to use a Java function.

When I say the other changes in #1186 are mainly ok, there are other new Java API methods that take Scala Partial Functions as inputs.

How do Java users even create a Scala Partial Function?

[He-Pin]

Yes, but a user can build a PartialFunction with the helper, such as the Receive builder.

[pjfanning]

Yes, but a user can build a PartialFunction with the helper, such as the Receive builder.

The ReceiveBuilder will not work for this stream-testkit use case. Unless we have a builder that can build a PartialFunction[Any, T] - does that exist?

[He-Pin]

@pjfanning Yes, org.apache.pekko.japi.pf.PFBuilder.

[pjfanning]

@pjfanning Yes, org.apache.pekko.japi.pf.PFBuilder.

Thanks - there seem to be some useful classes in the org.apache.pekko.japi.pf package.

https://pekko.apache.org/japi/pekko/current/org/apache/pekko/japi/pf/package-summary.html

[raboof]

@He-Pin what do you think we should do here? Should we document the issue we see in apache/pekko-persistence-cassandra#212 or is there a change we can make to stream-testkit to avoid the issue going forward?

If I understand correctly this is only an issue when using Scala 2.12, right? In that case I think the note in #1403 seems sufficient and IMHO we can close this issue (or keep it open for if anyone is interested in looking into it, but remove it from the 1.1.0 milestone)

[pjfanning]

I'll close this. It is documented and I think the consensus is not to fix it.