RFR 8207257 : Add VarHandle access test for value type declared in ref type

Roger Riggs Roger.Riggs at oracle.com
Thu Jun 13 19:57:40 UTC 2019

Hi Mandy,

Updated webrev:

On 6/13/19 3:12 PM, Mandy Chung wrote:
> On 6/13/19 8:22 AM, Roger Riggs wrote:
>> Please review an update to tests of VarHandles.  It adds support for 
>> an inline class
>> used in arrays and as a field of another inline class.
>> All tests pass except for those affected by 8207259[1] and the 
>> vh.compareAndSet methods.
>> The real changes are (mostly) to the templates, the other files are 
>> generated by the script.
>> The variable 'String' was changed to 'Class" so that it conditions 
>> all tests of Classes
>> including inline classes (not primitives)
>> Issue:
>> https://bugs.openjdk.java.net/browse/JDK-8207257
>> Webrev:
>> http://cr.openjdk.java.net/~rriggs/webrev-inlineclass-tests-8207257/index.html 
> test/jdk/java/lang/invoke/VarHandles/generate-vh-tests.sh
> + args="$args -KClass" Maybe "Object" instead of Class as the value is 
> either inline or indirect object.

That's ok, the distinction is between Object types and primitives.

> + * @bug 8156486 8207257 These repo-valhalla issues will not be used 
> when integrating to the main development repo and if you add it in 
> @bug, it would be clear when time comes to integration. I see no need 
> to change @bug.
ok, will remove bugid
>> [1]8207259 Ensure VarHandle compare and set/exchange work for value 
>> types
> I can't quickly spot the special cases or excluded test case for 8207259.
> Tried searching 8207259 in the new xxxPoint tests and can't find it
> either.
> Are they excluded?
No, I figured to wait until that bug was fixed to push.
It will avoid another round of updates later.

Thanks, Roger

> Mandy

More information about the valhalla-dev mailing list