CR 6899850 Updated, jeff.dinkins now responsible manager, P4 java/classes_util TESTBUG: DateRegression fails with "Date hashCode misbehaves"

[Eric Wang]

Hi Alan,

Please review the code changes 
http://dl.dropbox.com/u/90659131/fixes/6899850/webrev/java/util/Date/DateRegression.java.udiff.html
As described in monaco, this bug is caused by wrong assumption that 
precision of Date is second instead of millionsecond.
If you think OK, I'll send a formal request. This is a closed test, 
which group can review it?

Regards,
Eric

On 2012/8/18 15:22, Alan Bateman wrote:
>
> Eric - is this a test bug that could be added to your list? It seems 
> the chances of running into this issue is very low but intermittent 
> failures are killing us then maybe we should just fix this while it is 
> fresh in our minds.
>
> (Naoto - I should explain that Eric Wang is in the SQE team and has 
> been spending some time on fixing several of our regression tests that 
> we excluded, by putting down on ProblemList).
>
> -Alan
>
>
> On 16/08/2012 18:11, Alan Bateman wrote:
>> On 16/08/2012 18:06, Naoto Sato wrote:
>>> Hi Alan,
>>>
>>> Not directly related to this bug, but I am just curious. I've been 
>>> running JPRT jdk8 jobs with "-testset core" option, but have never 
>>> seen this failure. Is the setting for TL nightly build different 
>>> from JPRT's?
>>>
>>> Naoto
>>>
>> TL nightly runs the tests the same way as JPRT. I don't think we've 
>> seen it fail before, I guess we just got lucky.
>>
>> -Alan
>>
>