RFR: 8305875: Test TraceVirtualThreadLocals should be run with continuations only

Wed Apr 12 06:29:34 UTC 2023

On Wed, 12 Apr 2023 06:13:04 GMT, David Holmes <dholmes at openjdk.org> wrote:

> Sorry, have to ask, but why does the test fail without continuations? IIRC the emulated VTs use a carrier-per-VT so I would expect the ThreadLocal test to work just fine.

The debugging option is for the VirtualThread implementation only at this time, it's a no-op when running with -XX:+UnlockExperimentalVMOptions -XX:-VMContinuations. It could be made to work with the fallback implementation but it risks using thread locals recursively as formatters effectively run arbitrary code.

-------------

PR Comment: https://git.openjdk.org/jdk/pull/13436#issuecomment-1504728388