code review request: 7018928: test failure: sun/security/krb5/auto/SSL.java

[Weijun Wang]

On 02/15/2011 05:51 AM, Valerie (Yu-Ching) Peng wrote:
>
> So, the test failures are intermittent?

I think so.

> The changes look fine.

Thanks
Max

> Valerie
>
> On 02/13/11 11:49 PM, Weijun Wang wrote:
>> Hi Valerie
>>
>> http://cr.openjdk.java.net/~weijun/7018928/webrev.00/
>>
>> There was a failure of krb5 test SSL.java for jdk7b130 PIT. The
>> exception thrown is a java.net.NoRouteToHostException. The test uses
>> an internal DNS service and starts its own KDC and HTTPS server.
>>
>> I cannot replay the error so is not sure what the reason is. Anyway,
>> this kind of tests should definitely use run/othervm because it is
>> setting a number of system properties. I've added this argument to all
>> tests in the sun/security/krb5/auto directory that starts the internal
>> DNS service. Hopefully this would resolve the bug.
>>
>> I haven't made any changes to JGSS/krb5 codes and tests in the last
>> two week.
>>
>> Thanks
>> Max
>>
>>
>> -------- Original Message --------
>> *Change Request ID*: 7018928
>> *Synopsis*: test failure: sun/security/krb5/auto/SSL.java
>>
>>
>> === *Description*
>> ============================================================
>> New failure for jdk7b130 JPRT test on solaris-sparc:
>>
>> sun/security/krb5/auto/SSL.java - `main' threw exception:
>> java.net.NoRouteToHostException: Cannot assign requested address
>>
>>
>> === *Evaluation*
>> =============================================================
>> This test sets a number of system properties, and should use othervm
>> when running.
>>
>> Some of NameService tests (as part of CR 7012768) are changed.
>> Although none of the source changes should effect this test, but maybe
>> moving the tests causes them to now run in a different order, which of
>> course effects others in samevm.
>>
>