S RFR ASAP: 8026185: 8009130 broke GetMethodDeclaringClass/declcls001

[Karen Kinnear]

Thank you Coleen, Yumin and Harold. Coleen's fix is much better - I am just blind today.

I will make the one line change, rebuild, retest and send it out.

thanks Coleen!!!

Karen

On Oct 9, 2013, at 4:23 PM, yumin.qi at oracle.com wrote:

> Looks good.
> 
> As Coleen pointed out,  if default_methods() returns NULL, 'm' will be junk.
> 
> Either way is OK to me.
> 
> Thanks
> Yumin
> 
> On 10/9/2013 12:47 PM, Karen Kinnear wrote:
>> Folks,
>> 
>> webrev: http://cr.openjdk.java.net/~acorn/8026185/webrev/
>> bug: https://bugs.openjdk.java.net/browse/JDK-8026185
>> 
>> After all the code reviews for 8009130, Monday morning I did a small amount of refactoring and messed that up.
>> It broke in hotspot-rt nightly, so I need to get the fix in TODAY.
>> 
>> The code changes here go back to the code that Yumin, Serguei and Coleen actually did review
>> and which was extensively tested.
>> 
>> Testing (in progress)
>> linux:
>> nsk/jvmti/GetMethodDeclaringClass/declcls001 (002, 003)
>> jck api/java_tuil/TreeSet/ParallelStream.html (dependencies.cpp use case)
>> vm.quick.testlist
>> 
>>   which includes nsk/regression/b4226863 and nsk/stress/jck12a/jck12a009
>> Solaris:
>>   vm.quick-dtrace.testlist which includes
>>   8026217: dtrace/hotspot_jni/ALL/ALL001
>> 
>> thanks,
>> Karen
>> 
>> p.s. you can compare to these parts of the files you reviewed in
>> http://cr.openjdk.java.net/~acorn/8009130.4/webrev/
>> 
>> which incorporated comments from Lois, Yumin and Serguei and Coleen's naming change