Hi Roger, I think the test would be more robust if the hard-coded time out window was a function of the time factor the test was running under. Maybe 10 seconds * timeout IIRC, the timeout factor is available as an environment variable (or similar) in the running tests. I checked the jtreg docs quickly and didn't find it, but I believe the jdk regression tests have already had a pass to do this kind of update so there should be some guiding examples in the sources. Thanks, -Joe On 7/8/2015 1:16 PM, Roger Riggs wrote:
Ping and note this change is simplified after the fix for 8085981.
Please review this change to a ProcessHandle test to address intermittent failures. Unexpected subprocesses has been observed. The test succeeds if all of the processes under test are accounted for; others are ignored.
Webrev: http://cr.openjdk.java.net/~rriggs/webrev-exit-8130296/
Issue: https://bugs.openjdk.java.net/browse/JDK-8130296
Thanks, Roger