RFR (M): 8145148: InterfaceMethod CP entry pointing to a class should cause ICCE

[Coleen Phillimore]

On 5/16/16 9:21 PM, David Holmes wrote:
> Hi Coleen,
>
> On 17/05/2016 11:10 AM, Coleen Phillimore wrote:
>> Summary: pass tag in LinkInfo and check that resolved method type
>> matches tag, fix tests.
>>
>> Yumin's initial checkin of this code caused many test failures. These
>> have been addressed in this change and in previous changes to the
>> internal testbase tests.  The ASM change to enable passing boolean for
>> interface methods to the Handle constructor was checked in some time
>> ago.  This has been tested with our full nightly set.
>>
>> open webrev at http://cr.openjdk.java.net/~coleenp/8145148.01/webrev
>> bug link https://bugs.openjdk.java.net/browse/JDK-8145148
>
> Only an initial look so far. I don't understand why the AccessCheck 
> enum was introduced - it just seems to generate "noise" in converting 
> from booleans to AccessCheck ??

I had access_check be a bool but it was matching with constantTag passed 
in by mistake which cost me some time to debug.  So the type is safer, 
especially since they are default parameters.

Coleen

>
> Thanks,
> David
>
>> Thanks,
>> Coleen
>>
>>