Argh

[Stefan Ring]

On Thu, Nov 10, 2011 at 6:00 PM, Andrew Haley <aph at redhat.com> wrote:
> So, I'm trawling through jtreg failures to see which ones really
> are VM bugs,and I notice that some java.util.concurrent tests are
> failing, and even some really simple ones.  "Ah well," I thought,
> "there must be some problem with the memory barrier code."  So, I
> read through the source looking for the barriers.
>
> And that was the problem.  There are no barriers.  None, nada,
> zilch.  On ARM this is a really big problem: memory barriers are
> essential.

CACAO also doesn't have any for ARM yet, IIRC - don't feel bad ;).