RFR: Lambda: 8009130 JCK lambda test fails with IllegalAccessException
Karen Kinnear
karen.kinnear at oracle.com
Sun Oct 6 18:42:30 PDT 2013
Coleen,
Thank you so much for going over this and for the phone conversation. It helped me as well.
Details below,
thanks,
Karen
On Oct 6, 2013, at 7:18 PM, Coleen Phillmore wrote:
>
> On 10/6/2013 4:15 PM, Karen Kinnear wrote:
>>> 1169 if (m->method_holder()->is_interface() &&
>>> 1170 !m->is_abstract()) {
>> Done. Coleen will like this - thanks!
>
> Definitely! One of the benefits of being slow to the review.
Yes - this should help the long verbose debugging sections to be shorter and reduce the duplication.
>
> Karen, thank you for going over this code with me on the phone. I think it looks good. I have one request for a comment in InstanceKlass::deallocate_contents() where I you don't delete the methods in the default_methods array - ie. These methods are owned by super-classes and super-interfaces so don't delete them (or some such).
// Do NOT deallocate the default methods, they are owned by superinterfaces.
>
> Also the pattern in dependencies.cpp and jni.cpp is the same The difference between find_method and lookup_method in InstanceKlass is subtle. I think it would help to have a function with this pattern refactored as lookup_method_including_interfaces() with a comment that this looks in the default_methods array also. I searched through and don't see any other places where I think this might have been missed.
Thank you. Great idea. And linkResolver could use this too. I will do the refactoring in the morning so I make fewer careless mistakes
and send it to you.
Thank you for checking - missing that of course was a big concern to me.
>
> I don't have any additional comments. This does look good.
thank you very much,
Karen
> Thanks,
> Coleen
More information about the hotspot-runtime-dev
mailing list