RFR(XS): 8170919: LogStreamTest tests crash if they are run first

[Thomas Stüfe]

Thank you Kirill!

Kind Regards, Thomas

On Fri, Dec 9, 2016 at 9:33 AM, Kirill Zhaldybin <
kirill.zhaldybin at oracle.com> wrote:

> Thomas,
>
> Thank you for noticing this!
>
> I filed JDK-8170976 for this issue.
> I think we should fix it but since we do not support running gtests with
> external flags I do not think it is a big issue for now.
>
> Regards, Kirill
>
> On 08.12.2016 18:09, Thomas Stüfe wrote:
>
>> Hi,
>>
>> I wonder whether the test should use TEST_OTHER_VM?
>>
>> If I understand the coding in LogTestFixture correctly, it changes the
>> log configuration and does not restore it to its original value but to a
>> default state. So, logging would not be possible for follow up tests in
>> the same VM.
>>
>> So, test should either run in an own VM or LogTextFixture should
>> snapshot and restore the original log configuration.
>>
>> Kind Regards, Thomas
>>
>>
>> On Thu, Dec 8, 2016 at 3:41 PM, Kirill Zhaldybin
>> <kirill.zhaldybin at oracle.com <mailto:kirill.zhaldybin at oracle.com>> wrote:
>>
>>     Marcus,
>>
>>     Thank you for review!
>>
>>     Regards, Kirill
>>
>>     On 08.12.2016 17:08, Marcus Larsson wrote:
>>
>>         Hi,
>>
>>
>>         On 2016-12-08 14:32, Kirill Zhaldybin wrote:
>>
>>             Dear all,
>>
>>             Could you please review this fix for 8170919?
>>
>>             Since the tests use Resource Mark they should be run with
>>             TEST_VM not TEST.
>>
>>             WebRev:
>>             http://cr.openjdk.java.net/~kzhaldyb/webrevs/JDK-8170919/web
>> rev.00/
>>             <http://cr.openjdk.java.net/~kzhaldyb/webrevs/JDK-8170919/we
>> brev.00/>
>>
>>
>>         Looks good.
>>
>>         Marcus
>>
>>             CR: https://bugs.openjdk.java.net/browse/JDK-8170919
>>             <https://bugs.openjdk.java.net/browse/JDK-8170919>
>>
>>             Thank you.
>>
>>             Regards, Kirill
>>
>>
>>
>>
>>
>