[8u] RFR 8263601: TEST_BUG: remove ActivationLibrary.DestroyThread and have callers call rmid.destroy() instead

[Zhengyu Gu]

Hi,

Make sense, especially, 8032050 backport is clean.

Thanks for the suggestion.

-Zhengyu

On 4/27/21 4:26 PM, Hohensee, Paul wrote:
> Why not backport 8032050 first? It isn't very big, in fact most of it is waitFor().
> 
> Thanks,
> Paul
> 
> -----Original Message-----
> From: jdk8u-dev <jdk8u-dev-retn at openjdk.java.net> on behalf of Zhengyu Gu <zgu at redhat.com>
> Date: Monday, March 15, 2021 at 10:21 AM
> To: jdk8u-dev <jdk8u-dev at openjdk.java.net>
> Subject: [8u] RFR 8263601: TEST_BUG: remove ActivationLibrary.DestroyThread and have callers call rmid.destroy() instead
> 
> Hi,
> 
> I would like to backport this patch to 8u for parity with Oracle 8u301.
> 
> Original bug: https://bugs.openjdk.java.net/browse/JDK-8035000
> Original patch: http://hg.openjdk.java.net/jdk9/jdk9/jdk/rev/1b599b4755bd
> 
> The original patch applies cleanly.
> 
> However, patched RMID.java uses waitFor(long timeout) method, that was
> introduced by JDK-8032050 [1] and it is not backported to 8u. Therefore,
> I propose backport this specific function from JDK-8032050 with this patch.
> 
> 
> --- old/test/java/rmi/testlibrary/JavaVM.java   2021-03-15
> 13:03:32.298020591 -0400
> +++ new/test/java/rmi/testlibrary/JavaVM.java   2021-03-15
> 13:03:32.157020372 -0400
> @@ -26,6 +26,7 @@
>    import java.io.OutputStream;
>    import java.util.Arrays;
>    import java.util.StringTokenizer;
> +import java.util.concurrent.TimeoutException;
> 
>    /**
>     * RMI regression test utility class that uses Runtime.exec to spawn a
> @@ -183,6 +184,39 @@
>            errPipe.join();
>            return status;
>        }
> +    /**
> +     * Causes the current thread to wait the vm process to exit, if
> necessary,
> +     * wait until the vm process has terminated, or the specified
> waiting time
> +     * elapses. Release allocated input/output after vm process has
> terminated.
> +     * @param timeout the maximum milliseconds to wait.
> +     * @return exit value for vm process.
> +     * @throws InterruptedException if the current thread is interrupted
> +     *         while waiting.
> +     * @throws TimeoutException if subprocess does not end after timeout
> +     *         milliseconds passed
> +     */
> +     public int waitFor(long timeout)
> +            throws InterruptedException, TimeoutException {
> +         if (vm == null)
> +             throw new IllegalStateException("can't wait for JavaVM
> that isn't running");
> +         long startTime = System.currentTimeMillis();
> +         long rem = timeout;
> +
> +         do {
> +            try {
> +                int status = vm.exitValue();
> +                outPipe.join();
> +                errPipe.join();
> +                return status;
> +            } catch (IllegalThreadStateException ex) {
> +                if (rem > 0) {
> +                    Thread.sleep(Math.min(rem, 100));
> +                }
> +            }
> +            rem = timeout - (System.currentTimeMillis() - startTime);
> +         } while (rem > 0);
> +         throw new TimeoutException();
> +     }
> 
>        /**
>         * Starts the subprocess, waits for it to exit, and returns its
> exit status.
> 
> 
> Thanks,
> 
> -Zhengyu
> 
> [1] https://bugs.openjdk.java.net/browse/JDK-8032050
> 
>