RFR(M): 8160299 - Test8015436 doesn't check which method was executed

[dmitrij pochepko]

I'm trying to follow unified approach. We have a number of tests 
following both schemes: few testclasses launch themselves in different 
modes(driver and test) and a number of tests has separate driver. I 
remember SQE locally discussed that separate launcher/driver file is 
preferrable to improve readability(also, in case we have exception in 
test code, sometimes it takes time to understand which vm stacktrace 
belongs to) and so an approach with separate driver prevails.

Thanks,
Dmitrij

On 08.08.2016 21:24, Vladimir Kozlov wrote:
> On 8/8/16 11:22 AM, dmitrij pochepko wrote:
>> Thank you for attentive review.
>>
>>
>> oa = ProcessTools.executeTestJvm(Test8015436.class.getName());
>>
>> this call forks another vm, wait for exit and provide output for 
>> analyzing.
>> So, basically Driver8015436 is just a wrapper for original Test8015436
>
> Yes, I know that. I just asked was it easier to do than modifying the 
> test itself and do not add an other file?
>
> Thanks,
> Vladimir
>
>>
>> Thanks,
>> Dmitrij
>>
>>
>> On 08.08.2016 20:45, Vladimir Kozlov wrote:
>>> Changes seem fine. Why not use fork and look on result of forked VM?
>>>
>>> Thanks,
>>> Vladimir
>>>
>>> On 8/8/16 6:37 AM, Dmitrij Pochepko wrote:
>>>> Hi,
>>>>
>>>> please review fix for 8160299 - Test8015436 doesn't check which method
>>>> was executed
>>>>
>>>> Test wasn't strict enough and test output which reflects main test 
>>>> logic
>>>> was unchecked. This fix adds simple driver class which performs
>>>> respective output checks.
>>>>
>>>> webrev: http://cr.openjdk.java.net/~dpochepk/8160299/webrev.01/
>>>> CR: https://bugs.openjdk.java.net/browse/JDK-8160299
>>>>
>>>> I've tested this fix on linux-amd64.
>>>>
>>>> Thanks,
>>>> Dmitrij
>>