On 12/10/2014 11:52 AM, Staffan Larsen wrote:
Looks good. Perhaps use RuntimeMXBean.getInputArguments() instead of the dcmd?
Sounds good. I will change the test to use
RuntimeMXBean.getInputArguments() then.
Thanks!
-JB-
/Staffan
On 10 dec 2014, at 11:40, Jaroslav Bachorik <jaroslav.bacho...@oracle.com>
wrote:
A kindly reminder ...
On 12/04/2014 10:42 AM, Jaroslav Bachorik wrote:
Please, review the following test change
Issue : https://bugs.openjdk.java.net/browse/JDK-8034263
Webrev: http://cr.openjdk.java.net/~jbachorik/8034263/webrev.00
The test fails intermittently only for CMS collector -
"One thing about CMS is that it does concurrent work. So just a theory
would be that a concurrent cycle is started and sometime finishes (which
might trigger a notification) and sometimes does not finish before the
test has completed. " (from discussion with Bengt Rutisson)
The 'fix' is to relax the condition for when CMS is used and check that
at least the expected number of notifications was triggered.
Thanks,
-JB-
