RFR: 8146665: Test jdk/test/java/util/logging/LogManager/Configuration/updateConfiguration/UpdateConfigurationTest.java fails - missing expected output

Daniel Fuchs daniel.fuchs at oracle.com
Sat Jan 9 15:40:27 UTC 2016


Hi Mandy,

On 1/8/16 8:51 PM, Mandy Chung wrote:
>> On Jan 8, 2016, at 7:36 AM, Daniel Fuchs<daniel.fuchs at oracle.com>  wrote:
>> >
>> >Hi,
>> >
>> >Please find below a tentative fix for:
>> >
>> >8146665: Test jdk/test/java/util/logging/LogManager/Configuration/updateConfiguration/UpdateConfigurationTest.java fails - missing expected output
>> >https://bugs.openjdk.java.net/browse/JDK-8146665
>> >
>> >Webrev:
>> >http://cr.openjdk.java.net/~dfuchs/webrev_8146665/webrev.00/
>> >
>> >The fix increases the numbers of gc performed (which may or may
>> >not fix the issue), but more importantly fixes the finally { } clause
>> >of the test(...) method, where a failure in releasing the second
>> >logger reference may be hiding the original failure.
> line 280 adding the suppressed == null check is okay.
> Is the additional GC needed?

It's probably better to call gc() at least once before calling poll().

>  Are you able to reproduce the failure?

No - I have tried to run the test with the same set of options
on my local machine but never managed to make it fail.

The suppressed == null check should make it possible to see the
actual failure next time it fails, if it fails again.
Maybe that will be more informative.

best regards,

-- daniel





More information about the core-libs-dev mailing list