RFR(S) 8223050: JVMCI: findUniqueConcreteMethod() should not use Dependencies::find_unique_concrete_method() for non-virtual methods

[dean.long at oracle.com]

Thanks Doug!

dl

On 6/7/19 1:13 AM, Doug Simon wrote:
> Change looks good to me.
>
> -Doug
>
>> On 7 Jun 2019, at 04:41, Vladimir Kozlov <vladimir.kozlov at oracle.com> wrote:
>>
>> Seems reasonable to me but Graal experts should review it.
>>
>> Thanks,
>> Vladimir
>>
>> PS: please, run more graal tiers - I added comment in bug report.
>>
>> On 6/6/19 7:33 PM, dean.long at oracle.com wrote:
>>> https://bugs.openjdk.java.net/browse/JDK-8223050
>>> http://cr.openjdk.java.net/~dlong/8223050/webrev/
>>> Avoid generating unnecessary dependencies for methods that can be statically bound.  While working on this I discovered that Graal's HotSpotResolvedJavaMethodImpl.canBeStaticallyBound() doesn't match that of the VM, because it was missing a check for <init>.
>>> dl