RFR(S) : 8243432 : use reproducible random in :vmTestbase_vm_defmeth

[David Holmes]

On 1/05/2020 1:38 pm, Igor Ignatyev wrote:
> Hi David,
> 
> one reason is to remove redundant code, vmTestbase have several classes which provide means to specify seed, and it might have made sense in old test-harness where each testlist/testgroup might have been seen as a separate product, it's not really a case anymore.
> 
> another is to unify how we specify a seed, so it would be easier to have a seed value "stable" for a build/CI run, which will help to compare results of "randomness" tests from different runs w/o their re-execution, right now if you have such test failing let's say on windows-x64 and linux-x64, you can't tell if it's platform specific or not until you rerun on one of the platforms w/ the seed from another. and from that perspective, 8243432  and other subtasks of 8241623 are just preparation step, yet w/ an extra value of their own.

So can't we program the obtaining of the seed into the existing 
framework? What if you just switch the default seed from 0 (random) to 
the mystical 42?

David

> Thanks,
> -- Igor
>   
>> On Apr 30, 2020, at 7:48 PM, David Holmes <david.holmes at oracle.com> wrote:
>>
>> Hi Igor,
>>
>> On 1/05/2020 3:23 am, Igor Ignatyev wrote:
>>> http://cr.openjdk.java.net/~iignatyev/8243432/webrev.00
>>>> 22 lines changed: 3 ins; 12 del; 7 mod;
>>> Hi all,
>>> could you please review this small patch?
>>> from JBS:
>>>> this subtask is to use j.t.l.Utils.getRandomInstance() as a random number generator, where applicable, in : vmTestbase_vm_defmeth test group and marking the tests which make use of "randomness" with a proper k/w.
>>
>> Sorry I don't like this change at all. The code already supports using a deterministic seed so why not just set it?
>>
>> Thanks,
>> David
>>
>>> testing: : vmTestbase_vm_defmeth test group
>>> JBS: https://bugs.openjdk.java.net/browse/JDK-8243432
>>> webrevs:
>>> - code changes: http://cr.openjdk.java.net/~iignatyev//8243432/webrev.00.code
>>>> 16 lines changed: 1 ins; 12 del; 3 mod;
>>> - adding k/w: http://cr.openjdk.java.net/~iignatyev//8243432/webrev.00.kw
>>>> 2 lines changed: 2 ins; 0 del; 0 mod;
>>> - full: http://cr.openjdk.java.net/~iignatyev//8243432/webrev.00
>>>> 22 lines changed: 3 ins; 12 del; 7 mod;
>>> Thanks,
>>> -- Igor
>