Dr. Christian Kohlschütter

The latest 2.10.0-SNAPSHOT should remedy this situation. Threads should no longer block each other if both the client and server run on VirtualThreads, and the test code provided seems to...

@urbanchef Should be ready in a few days, sorting out a couple of remaining issues first. Please try the latest SNAPSHOT version and let me know if you find any...

junixsocket 2.10.0 has been released. Please verify and re-open if necessary. Thanks again for reporting , @cenodis !

Thanks for reporting, @cenodis! Please verify against the current snapshot (changes referenced above). If there are still discrepancies between junixsocket and vanilla Java, please provide some code/unit test to demonstrate,...

@cenodis Thanks for providing these additional details! I've added some more changes to make your tests pass, apart from the aforementioned interrupt flag state — and ClosedChannelException vs. ClosedByInterruptException: ClosedByInterruptException...

Please try again with the latest snapshot (2.10.0-20240630.191437-9 corresponding to commit 7025a0495d96770d4d1ba76b986fc824518e786e). I've reworked both the exception handling code as well as your unit test, which is now also included...

OK, so I think I have it now. There are a few things I was able to change, and now we get the expected `ClosedByInterruptException` in all cases where we...

Please verify with code from the latest commit 9f5377f88a3b3138e22690d04758bfd34e88fb59. You can also run the latest [selftest jar](https://oss.sonatype.org/content/repositories/snapshots/com/kohlschutter/junixsocket/junixsocket-selftest/2.10.0-SNAPSHOT/junixsocket-selftest-2.10.0-20240701.224420-13-jar-with-dependencies.jar) as follows: ``` java \ -Dselftest.issue.158.debug=true \ -Dselftest.enable-module.junixsocket-common.JEP380=true \ -Dselftest.enable-module.junixsocket-common.JavaInet=true \ -jar junixsocket-selftest-2.10.0-20240701.224420-13-jar-with-dependencies.jar...

Java bug reported to Oracle, [JDK-8335600](https://bugs.openjdk.org/browse/JDK-8335600)

@cenodis I'm glad we got this resolved. Investigating this bug was quite fruitful, as we found a couple more issues in junixsocket, plus a JDK bug :)