He-Pin commented on issue #2904: URL: https://github.com/apache/pekko/issues/2904#issuecomment-5010665995
Closing as resolved by #2999 (commit 44e2aaecaed37d03f0f82d736cad2e885e370ff7). I checked the complete log from the original failing nightly run: https://github.com/apache/pekko/actions/runs/24891032425. The first failure was InputStreamSourceTest stochastic_spec103_mustSignalOnMethodsSequentially at iteration 43: it expected completion after 10 elements but received 0. The actor-system-terminate timeout reported in this issue was secondary; the remaining graph was flow-69-0-take. The old TCK publisher used an infinite CPU-busy InputStream, the default 8192-byte chunk size, and take(elements), while the TCK counts publisher elements rather than bytes. #2999 replaced it with a finite InputStream, chunkSize = 1, and removed take(elements), so the publisher now completes deterministically and its element count matches the TCK contract. Verification: - main at 1fa250d450e18bbebd0374378b6e2a8069473e46 - Oracle JDK 25.0.2, Scala 3.3.8, 4 available CPUs, and the current JDK 25 nightly virtualize/minimum-runnable settings - InputStreamSourceTest: 38 total TCK cases, 26 passed, 12 skipped, 0 failed; 1 suite completed, 0 aborted - After #2999 merged, 18 nightly workflows from June 25 through July 13 completed successfully. The only overall failure in that interval was the unrelated MixedProtocolClusterSpec on July 10; its JDK 25 / Scala 3.3.x job succeeded and InputStreamSourceTest passed. -- 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]
