On Tue, 3 Jan 2023 17:43:43 GMT, Volker Simonis <simo...@openjdk.org> wrote:
>> Looks good to me. Thanks! > >> Looks good to me. Thanks! > > Thanks @XueleiFan! > > I've updated the copyright year to 2023 and will wait one or two more days > just in case @ascarpino wants to take one more look as well. > Hi @simonis, I am sorry for chiming in so late on this issue. I do think it > might be worthwhile to make your proof-of-concept code into a jtreg test as > you mentioned in your summary. I think it really comes down to how feasible > the conversion would be. It's always better to have an automated test if we > can, but it depends on if jtreg code can get access to reliable information > about the session tickets via the session cache and know that things are > behaving as intended. I don't think the current reproducer can easily be converted into a cheap and stable jterg test. The current version is quite heavyweight because it calls `jcmd GC.class_histogram` and not really stable because it depends on the number of live `StatelessKey` objects in the heap. So for now I'd prefer to finally fix this issue even without an attached automatic test. ------------- PR: https://git.openjdk.org/jdk/pull/11590
