On Fri, 10 Jul 2026 04:33:23 GMT, Serguei Spitsyn <[email protected]> wrote:
>Since native method stopThread() ignores JVMTI_ERROR_OPAQUE_FRAME, the test >can pass even if every StopThread call returns OPAQUE_FRAME and no async >exception is ever delivered. But I'm not sure what can be done here to ensure >the async exception is delivered and also keep the test stable. > The main purpose of the test is to verify that we don’t install the async exception at a critical point where it could cause a crash, so that scenario is fine (we just found each target at an unsafe place and rejected the request). > Would it make sense to count the OPAQUE_FRAME errors, so sum of the number of > threads with async exceptions and with OPAQUE_FRAME is equal to total number > of the test threads? > I updated the code to track the cases where the async exception is installed to match it against the thrown cases. Tracking also `OPAQUE_FRAME` to match both against sum of all threads should not be necessary as any other error would cause `check_jvmti_status` to exit with fatal error. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/31759#discussion_r3571993403
