code review request: 7170308: timing error in the krb5 test SSL.java

Weijun Wang weijun.wang at oracle.com
Mon May 21 05:43:00 UTC 2012


How about this?

  http://cr.openjdk.java.net/~weijun/7170308/webrev.01/

The flag is now a state.

Thanks
Max

On 05/21/2012 11:23 AM, Xuelei Fan wrote:
> A minor comment.
>
> There may be a potential dead loop in client/main block. If the server
> runs into exception before set the serverReady, the client will deadly
> wait for the setting.
>
> I think it may be a little safer to have the set of serverReady in a
> final clause.
>
> try {
>      SSLServerSocketFactory sslssf =
>      ...
> } finally {
>      serverReady = true;
> }
>
> Xuelei
>
> On 5/21/2012 10:52 AM, Weijun Wang wrote:
>> Hi Xuelei
>>
>> I've adopted your style of synchronizing the SSL client and server.
>> Please take a review at
>>
>>    http://cr.openjdk.java.net/~weijun/7170308/webrev.00/
>>
>> Noreg-self.
>>
>> Thanks
>> Max
>>
>> On 05/21/2012 10:14 AM, weijun.wang at oracle.com wrote:
>>> *Change Request ID*: 7170308
>>> *Synopsis*: timing error in the krb5 test SSL.java
>>>
>>> === *Description*
>>> ============================================================
>>> In the test the client side waits 2 seconds for the server to start,
>>> but on some machines this is still not enough. We should using a flag
>>> to synchronize their actions.
>>>
>



More information about the security-dev mailing list