On Fri, 18 Oct 2024 19:31:27 GMT, Harshitha Onkar <hon...@openjdk.org> wrote:
>> Indeed, the test does not play nicely with timeout in `PassFailJFrame`. >> >> The workaround could be to pause the timeout. In addition to that, we could >> set the timeout to 15 minutes. If the user changes time not -1 / +2 hours >> but rather -5 / +10 minutes, it should be enough. >> >>> What an .. interesting .. test. No idea what this does to the jtreg test >>> harness too. >> >> The jtreg test harness turns off its timeout for manual tests, as far as I >> understand. By the time, the test finishes, the time on the system should be >> correct again. >> >>> FWIW I think this test should be deleted rather than trying to make it >>> nicer. >> >> Another option could be to mark the test with `@ignore` so that it's not run >> regularly. The test could still be useful; even though changing time isn't >> common… >> >> I wonder if going to/from DST reproduces the problem in >> [JDK-7096375](https://bugs.openjdk.org/browse/JDK-7096375) for which the >> test was written. > >> The workaround could be to pause the timeout. In addition to that, we could >> set the timeout to 15 minutes. If the user changes time not -1 / +2 hours >> but rather -5 / +10 minutes, it should be enough. > > The workaround sounds good and may be also include pause timer in the test > instructions as below. > > > Pause the timer and reduce the system time to 10 mins less than current time. > ...... > ..... > Pause the timer and increase the system time by adding 10 mins to the current > time. > > > Yes I agree, retaining the test might be a good idea since it was created for > - > [JDK-7096375](https://bugs.openjdk.org/browse/JDK-7096375). Sounds good for a workaround. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/21476#discussion_r1808299946
