On Mon, 1 Aug 2022 14:21:53 GMT, Maxim Kartashev <[email protected]> wrote:
> On the issue of the second test, I still believe it is useful: > > * it demonstrates a legitimate problem with `Disposer`, albeit admittedly a > different one than reported in the bug, > * it shows the problem more clearly as the problematic usage pattern of > `Disposer` is contained in the test itself and not buried in JDK code, > * it crashes more reliably, > * it shows that the problem (at least potentially) exists on Linux also, > * it does not have to be marked headful. I think it might help us ensure that such a problem will not actually creep in, but I think it is testing a scenario that is not the cause of the crash etc. However I suppose in the end all our passing tests are testing things that aren't a problem right now ! .. so OK .. ------------- PR: https://git.openjdk.org/jdk/pull/9362
