https://bz.apache.org/bugzilla/show_bug.cgi?id=64733
--- Comment #18 from Jaikiran Pai <jaiki...@apache.org> --- > After the deadlock fix it does not affect our tests (except we have this > stack traces in the log, but that's not the biggest problem). That's good to hear. That way we can proceed with the Ant release, if necessary, even if this IOException issue isn't solved in the meantime. > Would a simple testcase where i can reproduce the "IOException: Pipe broken" > help you? Yes that will definitely help. Although, looking at the JDK code which deals with the PipedInputStream/PipedOutputStream, I think I see what the issue could be - the way the implementation of those classes decides about the "writerThread" can lead to issues like these where the tests could be run in different threads and each one of them writing to System.out/System.err although not concurrently. I'm not too sure I can fix this soon, but at least having a testcase which reproduces this will help us. -- You are receiving this mail because: You are the assignee for the bug.
