He-Pin opened a new pull request, #2728: URL: https://github.com/apache/pekko/pull/2728
## Motivation JDK 25 introduced changes to `ForkJoinPool` scheduling behavior that cause ~100+ stream test failures per nightly CI run. The most affected tests are high-throughput stream tests (`FlowFlatMapConcatParallelismSpec`, `HubSpec`, `MapAsyncPartitionedSpec`, `AggregateWithBoundarySpec`) that have timing-sensitive assertions. ## Modification 1. **Nightly CI (`.github/workflows/nightly-builds.yml`)**: Added `timefactor=3` for JDK 25+ builds, scaling all `dilated` timeouts across the entire test suite 2. **`FlowFlatMapConcatParallelismSpec`**: Added explicit `PatienceConfig(30.seconds)` for high-element-count tests (100K elements) 3. **`HubSpec`**: Added explicit `PatienceConfig(30.seconds)` for 20K+ element throughput tests 4. **`MapAsyncPartitionedSpec`**: Increased patience from 5s to 15s (3x, matching timefactor) 5. **`AggregateWithBoundarySpec`**: Increased `Await.result` timeouts from 10s to 30s (3x) ## Result - All modified tests pass locally - The `timefactor=3` in nightly CI provides a safety net for tests without explicit overrides - Explicit timeout increases target the highest-failure tests identified from nightly CI logs ## References - Fixes https://github.com/apache/pekko/issues/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]
