hg: jdk9/dev/jdk: 8030690: TEST_BUG java/nio/Buffer/Chars.java fails intermittently

[Alan Bateman]

On 20/12/2013 14:16, roger riggs wrote:
> This a case where to make the test reproducable, the test should have 
> a fixed
> randomized seed.  Or at least report the seed being used in the test.
For the test proper then it logs the buffer type, position and limit 
prior to each test so that there are crumbs in the event that it fails. 
It was just unfortunate that there was a bug in the setup. I don't think 
we've actually seen this, it's just that we are chasing a more important 
issue. That important issue being a stack overflow in the reporting by 
TestNG after the test completes successfully. Not clear yet whether this 
is a hotspot, TestNG, a char decoder or something else.

-Alan.