On 17/08/2013 14:06, Peter Levart wrote:
Hi,
Please review the fix for a test that tries to provoke a deadlock when
parsing annotations:
http://cr.openjdk.java.net/~plevart/jdk8-tl/AnnotationType/fix1/webrev.01/
The test was failing on slow machines when run with -Xcomp due to a short
time out (5s). The fix makes the test only fail when the dead-lock
situation is detected - never times out. I also re-worked the dead-lock
detection so that it is more correct now (using ThreadMXBean).
I remember an early version of the test used 500ms and I suggested you
bump it to 5000ms on the assumption that it would be sufficient. It just
demonstrates again that timeouts almost always bite back. I've looked at
the webrev and it looks good to me, thanks again for picking this one up.
-Alan
