Thanks Paul!
David
On 8/05/2018 3:01 AM, Paul Sandoz wrote:
I am not too familiar with the overall test but i could follow the logic for
the additions, looks good to me.
Paul.
On May 6, 2018, at 3:11 PM, David Holmes <[email protected]> wrote:
webrev: http://cr.openjdk.java.net/~dholmes/8202686/webrev/
bug: https://bugs.openjdk.java.net/browse/JDK-8202686
JDK-8200167 added additional checks to ensure receiver typechecks were in place
where needed for invokespecial invocations. There was one variation missing in
the test: invoking a final Object method via an interface reference. This is
potentially different to a non-final Object method because the finality means
the method is treated as non-virtual and so activates different call paths in
the logic.
No unexpected behaviour was found but the test case is added for completeness
(which is significant when we adapt this invokeinterface testing for the
nestmates work (JEP 181).
There were also a couple of positive MH invocations missing from the test so
they were added for completeness as well.
Thanks,
David
