On Wed, 21 Jan 2026 21:06:44 GMT, Phil Race <[email protected]> wrote:
>A hang one in 100 times doesn't sound like much until you know that it can get >run 20 times a day .. so every few days you have a failure. Those numbers >aren't exact but are in the ball park. I do not agree with this approach. Adding the sound keyword to a test that is intended to run on systems with and without audio is the wrong solution. It does not address the underlying problem. Instead, it hides the issue and reduces test coverage. The hang in the "com.sun.media.sound.DirectAudioDevice.nOpen" seems like a generic code path, so it will not help. ------------- PR Comment: https://git.openjdk.org/jdk/pull/29353#issuecomment-3786609620
