[concurrency-interest] Difficulty interpreting jcstress results

Aleksey Shipilev aleksey.shipilev at oracle.com
Thu Aug 22 08:02:26 EDT 2013


This seems to indicate the fatal error, not the test failure. Maybe the VM crashed after a while? The forked VM can still have communicated a few results before the crash. Let me see if I can get more reliable reporting in these cases.

In the mean time, can you try to disable forking with "-f 0" and re-run the test selectively with "-t .*FloatBuffer.*"?

-Aleksey.

On 22.08.2013, at 13:43, Andrew Haley <aph at redhat.com> wrote:

> I'm getting an odd result:
> 
> FloatBufferAtomicityTests$FloatTest               ERROR    Error running the test
> 
> But when I open the test
> 
> Observed state    Occurrence    Expectation    Interpretation    Refs
> [-1]    0    ACCEPTABLE    Seeing the complete update.    
> [0]    484    ACCEPTABLE    Seeing the default value, this is valid under race.    
> [-1082130432]    456    KNOWN_ACCEPTABLE    All other cases are unexpected, but legal
> 
> So, if it's legal, why is there an error?
> 
> Andrew.



More information about the Concurrency-interest mailing list