Reports that this test is failing intermittently over past few months.
It's a rare occurrence but I'd like to take steps to correct it.
I've removed the dependence on randomness in the bug.
I've fixed up the zip file creation logic to produce a real zip file
I've renamed the file to a unique file name per test run (and recorded
how long it takes to delete it)
Test now logs any exception encountered during zip entry read
I've verified that the test still fails and passes as expected.
https://bugs.openjdk.java.net/browse/JDK-8213561
http://cr.openjdk.java.net/~coffeys/webrev.8213561/webrev/
--
Regards,
Sean.
