On Sun, 2 Aug 2026 22:48:51 GMT, David Holmes <[email protected]> wrote:

>> I rewrote this to loop 100 times instead then reporting failure if it 
>> doesn't pass within 100 times. I hope this test won't be cursed on slower 
>> systems or windows.
>
> The "loop forever" version would trigger the timeout handler and then we 
> would get a core etc and be able to see why it has not completed. Now we will 
> just fail but we won't know why.

I see the point.  The loop forever version will give a core file or an mdmp 
file.  I don't know how else to account for less reliable systems that we might 
run this test on.

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/32092#discussion_r3706718471

Reply via email to