On Fri, 24 Apr 2026 05:40:43 GMT, Dean Long <[email protected]> wrote:

> Perhaps an alternative approach would be to treat a terminated platform 
> thread similar to how we treat an unmounted virtual thread: allow the current 
> thread to perform the handshake (passing nullptr for the Thread*) and extract 
> information from the Java object.
>
The target could change its state to terminated while we are in 
`Handshake::execute` already, so we would have to change 
`HandshakeOperation::do_handshake` for that, but that would affect all 
handshakes. If the thread already terminated I think there is no point in 
forcing execution of the closure since we won’t get any information.

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

PR Comment: https://git.openjdk.org/jdk/pull/30903#issuecomment-4329787919

Reply via email to