RFR 9: 8152005: sun/misc/SunMiscSignalTest.java failed intermittently
Please review this improvement to an intermittently failing test of sun.misc.Signal. The timeout value is increased in case of a busy system and for additional debug value the raising of the signal is retried. Webrev: http://cr.openjdk.java.net/~rriggs/webrev-signal-8152005/ Issue: https://bugs.openjdk.java.net/browse/JDK-8152005 Thanks, Roger
Hi Roger, If the test is only rarely failing as-is, I'd prefer to keep the first timeout short and only fallback to the longer timeout in the fallback case. What do you think? Thanks, -Joe On 3/28/2016 11:22 AM, Roger Riggs wrote:
Please review this improvement to an intermittently failing test of sun.misc.Signal. The timeout value is increased in case of a busy system and for additional debug value the raising of the signal is retried.
Webrev: http://cr.openjdk.java.net/~rriggs/webrev-signal-8152005/
Issue: https://bugs.openjdk.java.net/browse/JDK-8152005
Thanks, Roger
Hi Joe, ok, my thought was that extending the time out didn't affect the running time of the test and since the signal is handled by a separate thread a busy system might be the cause. I removed the re-raise of the signal so the test will fail if the signal is not received even after the longer timout. Webrev updated in place, http://cr.openjdk.java.net/~rriggs/webrev-signal-8152005/ Thanks, Roger On 3/28/2016 8:29 PM, Joseph D. Darcy wrote:
Hi Roger,
If the test is only rarely failing as-is, I'd prefer to keep the first timeout short and only fallback to the longer timeout in the fallback case.
What do you think?
Thanks,
-Joe
On 3/28/2016 11:22 AM, Roger Riggs wrote:
Please review this improvement to an intermittently failing test of sun.misc.Signal. The timeout value is increased in case of a busy system and for additional debug value the raising of the signal is retried.
Webrev: http://cr.openjdk.java.net/~rriggs/webrev-signal-8152005/
Issue: https://bugs.openjdk.java.net/browse/JDK-8152005
Thanks, Roger
Hi Roger, It wasn't entirely clear to me from the docs if extending the time would impact the expected running time of the text on an otherwise unloaded system. The current version looks fine. Thanks, -Joe On 3/29/2016 8:24 AM, Roger Riggs wrote:
Hi Joe,
ok, my thought was that extending the time out didn't affect the running time of the test and since the signal is handled by a separate thread a busy system might be the cause. I removed the re-raise of the signal so the test will fail if the signal is not received even after the longer timout.
Webrev updated in place, http://cr.openjdk.java.net/~rriggs/webrev-signal-8152005/
Thanks, Roger
On 3/28/2016 8:29 PM, Joseph D. Darcy wrote:
Hi Roger,
If the test is only rarely failing as-is, I'd prefer to keep the first timeout short and only fallback to the longer timeout in the fallback case.
What do you think?
Thanks,
-Joe
On 3/28/2016 11:22 AM, Roger Riggs wrote:
Please review this improvement to an intermittently failing test of sun.misc.Signal. The timeout value is increased in case of a busy system and for additional debug value the raising of the signal is retried.
Webrev: http://cr.openjdk.java.net/~rriggs/webrev-signal-8152005/
Issue: https://bugs.openjdk.java.net/browse/JDK-8152005
Thanks, Roger
participants (2)
-
Joseph D. Darcy
-
Roger Riggs