Hello, Audio Guru.
Please review the small fix for unstable test reported by mach5 in jdk9.
The test has two issues:
- It uses System.currentTimeMillis() which can be affected by the
system time change. Replaced by System.nanoTime().
- It maintains its own timeout(5 seconds) to catch the hang of the
test. It was found that on some overloaded system this period is not
enough(on the tested system was necessary to increase it to 30-40
seconds or it will fail a few times per 1000 executions). I removed this
internal timeout from the test(it is provided as the warning now), and
relies on jtreg default timeout. It has a benefit that when the test
hangs the jtreg will provide a jstack output which will be useful to
understand where is the problem.
Bug: https://bugs.openjdk.java.net/browse/JDK-8148915
Webrev can be found at: http://cr.openjdk.java.net/~serb/8148915/webrev.00/
--
Best regards, Sergey.
