RFR (S) JDK-8029415: java/lang/reflect/Method/invoke/TestPrivateInterfaceMethodReflect.java fails on all platforms with hs25-b61

[Coleen Phillimore]

On 12/4/2013 12:03 PM, Lois Foltan wrote:
>
> Thank you Karen & Harold for the review.  Harold, I concur about 
> 'is_reflect' as a good change for the future.

Noted. https://bugs.openjdk.java.net/browse/JDK-8029567

Coleen
>
> Lois
>
> On 12/4/2013 8:44 AM, harold seigel wrote:
>> Hi Lois,
>>
>> The change looks good.  Eventually the 'is_reflect' check could be 
>> put in its own method because it appears a few other places in the code.
>>
>> Harold
>>
>> On 12/4/2013 8:15 AM, Karen Kinnear wrote:
>>> Lois,
>>>
>>> Looks good. Many thanks for the discussion.
>>>
>>> thanks,
>>> Karen
>>>
>>> On Dec 3, 2013, at 8:39 PM, Lois Foltan wrote:
>>>
>>>> Please review the following fix:
>>>>
>>>> Webrev:
>>>>     http://cr.openjdk.java.net/~lfoltan/bug_jdk8029415/
>>>>
>>>> Bug: 
>>>> java/lang/reflect/Method/invoke/TestPrivateInterfaceMethodReflect.java 
>>>> fails on all platforms with hs25-b61
>>>>     https://bugs.openjdk.java.net/browse/JDK-8029415
>>>>
>>>> Summary of fix:
>>>>    Prior to emitting an ICCE during method resolution of an 
>>>> invokespecial that names a method in an
>>>>    indirect superinterface of the current class or interface, a 
>>>> check first must be made that the current
>>>>    class is not a dynamically-generated bytecode associated with 
>>>> the 1.4 reflection implementation.  This
>>>>    check added to LinkResolver::resolve_special_call() is patterned 
>>>> after the similar check in classfile/verifier.cpp's
>>>>    Verifier::is_eligible_for_verification().
>>>>
>>>> Tests:
>>>>    JCK vm & lang, JDK java.lang & java.util, vm.quick.testlist, 
>>>> hotspot/test/*, default method tests
>>
>