He-Pin opened a new pull request, #3004:
URL: https://github.com/apache/pekko/pull/3004
### Motivation
`GroupByTest` and `PrefixAndTailTest` abort the TCK suite on JDK 25 nightly
runs with:
```
java.lang.AssertionError: Failed in iteration 18 of 100.
Expected completion signal after signalling 10 elements (signalled 0),
yet did not receive it within 32000 ms
at
org.reactivestreams.tck.PublisherVerification.stochastic_spec103_mustSignalOnMethodsSequentially(...)
```
after roughly 18 iterations of
`stochastic_spec103_mustSignalOnMethodsSequentially`.
Both `createPublisher` implementations materialize an intermediate stream
and call `Await.result(future, 3.seconds)` to obtain the `Source` wrapped by
the test `Publisher`. The `3.seconds` value is a hardcoded constant that does
not honour `pekko.test.timefactor`, so the JDK 25 nightly's compounded GC +
ForkJoinPool pressure can push a single materialization beyond 3s during the
100-iteration stochastic test, which surfaces as a `Publisher` that signals
zero elements (and then aborts the whole TCK suite).
`Timeouts` already scales every other field by `timeFactor`.
`publisherShutdownTimeoutMillis` is the only existing field that wasn't scaled.
### Modification
- Add `Timeouts.materializerTimeoutMillis`, scaled by `timeFactor`.
- Route the `Await.result` calls in `GroupByTest` and `PrefixAndTailTest`
through the new helper instead of the hardcoded `3.seconds`.
- Apply the same `timeFactor` scaling to the existing
`publisherShutdownTimeoutMillis` so the whole `Timeouts` object behaves
consistently.
### Result
On local runs (`timeFactor=1`) the effective budget stays at 3s, so the test
still flags genuine materialization regressions. On the JDK 25 nightly
(`timeFactor=4`) the budget grows to 12s, which covers the observed worst-case
materialization delays without masking real bugs.
### Tests
- `sbt stream-tests-tck/Test/compile` (Scala 2.13) succeeded locally.
### References
- Failing nightly: https://github.com/apache/pekko/actions/runs/26571377673
- Failing nightly: https://github.com/apache/pekko/actions/runs/26611472055
- Tracking issue: #2573
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
