RFR: jsr166 jdk9 integration wave 2

[Martin Buchholz]

On Mon, Nov 23, 2015 at 8:04 PM, joe darcy <joe.darcy at oracle.com> wrote:

> Hello,
>
> There is a high barrier to consistent use of new jtreg keywords. FWIW, I
> would prefer expanding the scope of @key randomness to defining a new @key
> nondeterminstic.
>

It should come as no surprise to anyone that jtreg tests in j.u.concurrent
are usually nondeterministic.  When we have deterministic tests, they
usually end up as tck tests instead.   Do you really think adding @key
randomness to all of those tests will help?

Our approach instead has been to try to make these tests much more robust,
including the use of lots of noisy testing.