On Fri, 1 Sep 2023 14:58:56 GMT, Kevin Rushforth <k...@openjdk.org> wrote:
>> they become predictable when the seed is known. > > Yes, but then you would need to modify the test to rerun it with the same > seed as the time it failed. If we had a utility that made this easy, then > maybe, but I'd still be skeptical. I am not convinced that whatever benefit > we might get is worth the effort or the unpredictability. the main benefit is that every new run effectively extends the test coverage. however, in this particular case, the test won't be reproducible because it depends on many other factors like the platform load. ------------- PR Review Comment: https://git.openjdk.org/jfx/pull/1228#discussion_r1313162332
