On Tue, 18 Aug 2026 07:45:29 GMT, David Holmes <[email protected]> wrote:
>> The test support execution with any flags. So no need to make it flagless. >> I don't see now how it could be very useful right now, but not harmful as >> for the most of our tests. The standard practice is to don't restrict the >> execution in such cases. > > Well it is redundant to run it in a Xcomp batch as it already uses Xcomp. And > really this only cares about this specific compiled-code path so I see no > value running it with different GC, or virtual threads or ... That's might be a good idea to have something `vm.flags.specified` to mark the tests which don't make a sense with additional VM flags. So we can save some time by not running them to CI. However, we still might want to run them in ATR just to ensure that they don't find "unexpected" bugs. However, we want to differ this from `vm.flagless` which means that test doesn't support any external flags and might start failing. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/32363#discussion_r3814742248
