S RFR: Lambda: no superclass methods in defaults
Karen Kinnear
karen.kinnear at oracle.com
Mon Sep 16 14:59:50 PDT 2013
Thanks Yumin for understanding and for the quick review.
thanks,
Karen
On Sep 16, 2013, at 5:38 PM, Yumin Qi wrote:
> Karen,
>
> That is OK roll back the one change. In future, we may need to make all the 'defs' consistent.
>
> #ifdef _WINDOWS
> .....
> #else
>
> #endif // _WINDOWS
>
> Some time confusing, which part is for WINDOWS? Your style is better to pick up code though most of the 'defs' not in your style.
>
> Others look good.
> Please go ahead for integration.
>
> Thanks
> Yumin
>
> On 9/16/2013 2:27 PM, Karen Kinnear wrote:
>> Yumin,
>>
>> My apologies. I will put back the // ndef PRODUCT to match the other 15 lines.
>> This is a style issue and is quite prevalent in a lot of code in the code base. I actually prefer it,
>> but we can have that discussion when we review detailed style preferences.
>>
>> So I will put this one line back and let that be a future discussion.
>>
>> thanks,
>> Karen
>>
>> On Sep 16, 2013, at 5:14 PM, Yumin Qi wrote:
>>
>>> Hi, Karen
>>>
>>> Could you can change all the
>>>
>>> 259:#endif // ndef PRODUCT
>>> 461:#endif // ndef PRODUCT
>>> 585:#endif // ndef PRODUCT
>>> 631:#endif // ndef PRODUCT
>>> 748:#endif // ndef PRODUCT
>>> 762:#endif // ndef PRODUCT
>>> 770:#endif // ndef PRODUCT
>>> 778:#endif // ndef PRODUCT
>>> 904:#endif // ndef PRODUCT
>>> 912:#endif // ndef PRODUCT
>>> 956:#endif // ndef PRODUCT
>>> 972:#endif // ndef PRODUCT
>>> 987:#endif // ndef PRODUCT
>>> 1141:#endif // ndef PRODUCT
>>> 1165:#endif // ndef PRODUCT
>>>
>>> in this file?
>>>
>>> There was total 16, and now 15 after you changed one.
>>> (The line number based on my workspace, not accurate).
>>>
>>> Thanks
>>> Yumin
>>>
>>> On 9/16/2013 2:03 PM, Karen Kinnear wrote:
>>>> Updated webrev
>>>>
>>>> http://cr.openjdk.java.net/~acorn/8024647.02/webrev/
>>>>
>>>> Incorporating comments from Yumin and Keith. I would appreciate formal approval from Keith :-)
>>>>
>>>> Here is an example of what TraceMethodDefaults prints for this case:
>>>>
>>>> Class B requires default method processing
>>>> B
>>>> A
>>>> java/lang/Object
>>>> I
>>>> Slots that need filling:
>>>> m()V
>>>> Looking for default methods for slot m()V
>>>> Creating overpasses...
>>>> for slot: m()V
>>>> Selected method: A.m()V : in superclass
>>>> Created 0 overpass methods
>>>> Default method processing complete
>>>>
>>>> thanks,
>>>> Karen
>>>>
>>>> On Sep 13, 2013, at 11:24 PM, Karen Kinnear wrote:
>>>>
>>>>> Please review
>>>>>
>>>>> webrev: http://cr.openjdk.java.net/~acorn/8024647.01/webrev/
>>>>> https://bugs.openjdk.java.net/browse/JDK-8024647
>>>>>
>>>>> Bug: Default method resolution incorrectly handles private superclass methods.
>>>>> Problem was that during default method resolution, we are creating overpasses for
>>>>> superclass target methods. We only need to create these for superinterface target methods.
>>>>>
>>>>> Testing:
>>>>> test cases in the bug
>>>>> defmeth tests - no additional failures
>>>>> (these test cases should be added to the vm sqe defmeth tests)
>>>>>
>>>>> regression tests:
>>>>> java.util.streams - in progress
>>>>> jprt - in progress
>>>>> vm.quick.testlist - in progress
>>>>>
>>>>> thanks,
>>>>> Karen
>>>>>
>>>>>
>
More information about the hotspot-runtime-dev
mailing list