One more funny thing: lines 2543 and 2535 are calls to
System.arraycopy. But there is not arraycopy in the trace, so it's not clear
where out of bounds happens

2006/12/5, Mikhail Loenko <[EMAIL PROTECTED]>:
I tried to sort random arrays and tried to change elements of arrays
in another thread while sorting. Neither one failed. 100 000
iterations.

2006/12/5, Nathan Beyer <[EMAIL PROTECTED]>:
> I can say that I've observed some of those before. Specifically the
> index out of bounds exceptions thrown from Arrays.mergeSort called
> from ObjectStreamClass.computerSerialVersionUID.
>
> There's either a missing synchronization point in ObjectStreamClass or
> there's nasty bug in the Arrays.mergeSort method.
>
> -Nathan
>
> On 12/4/06, tatyana doubtsova <[EMAIL PROTECTED]> wrote:
> > I've reported http://issues.apache.org/jira/browse/HARMONY-2438 regarding
> > intermittent failures of swing module tests running in once mode on IBM VME.
> > Could anybody take a look at it.
> >
> > Thanks,
> > Tanya
> >
> >
>

Reply via email to