[13] RFR: 8219612: [TESTBUG] compiler.codecache.stress.Helper.TestCaseImpl can't be defined in different runtime package as its nest host

[Rahul Raghavan]

Thank you Igor.

I will continue to push the fix.
Also request help from Mandy to reply the backwards compatibility 
question correctly.

(understood from JBS notes, the IncompatibleClassChangeError type 
failure would have been generated for existing TestCaseImpl itself, if 
it's implementation had actual nestmate relationship validation !)


Thanks,
Rahul

On 25/03/19 11:14 PM, Igor Ignatyev wrote:
> Hi Rahul,
> 
> the fix looks good to me, I have a meta question though (most probably for Mandy): wouldn't this restriction break backwards compatibility?
> 
> Thanks,
> -- Igor
> 
>> On Mar 25, 2019, at 12:51 AM, Rahul Raghavan <rahul.v.raghavan at oracle.com> wrote:
>>
>> Hi,
>>
>> Please review the following fix proposal for JDK-8219612.
>> http://cr.openjdk.java.net/~rraghavan/8219612/webrev.00/
>>
>>
>> - https://bugs.openjdk.java.net/browse/JDK-8219612
>> - http://hg.openjdk.java.net/valhalla/valhalla/rev/ab7ea72963c9
>>
>> Thanks to Mandy Chung.
>> As mentioned in JBS above fix changeset is same as the one contributed by Mandy Chung, in valhalla repo nestmates branch.
>>
>> Testbug issue - compiler.codecache.stress.Helper.TestCaseImpl can't be defined in different runtime package as its nest host.
>>
>> Proposed fix - Test rewritten to use top-level classes rather then nested ones.
>>
>>
>> Thanks,
>> Rahul
>