On Wed, 26 Jun 2024 06:36:05 GMT, SendaoYan <s...@openjdk.org> wrote:
> Hi all, > The newly added test `com/sun/jdi/BreakpointOnClassPrepare.java` fails at > `SUSPEND_NONE` mode. > To make this test less noisy, should we disable the SUSPEND_NONE mode for now. We've done a lot of G1 testing without seeing this problem, and the likelihood of a GC with G1 right after the thread terminates seem like it would be very rare. I would expect this issue to possibly be somewhat common with a concurrent GC, but not with G1. Maybe there is something about your (automatically configured) default heap size that is causing a GC to be triggered at just the right time. ------------- PR Comment: https://git.openjdk.org/jdk/pull/19897#issuecomment-2192878042
