Backport 8259922: MethodHandles.collectArguments does not throw IAE

[babsingh]

Fix for https://bugs.openjdk.java.net/browse/JDK-8259922.

Fix: ibmruntimes/openj9-openjdk-jdk@ab6b9c3.

Signed-off-by: Babneet Singh sbabneet@ca.ibm.com

[babsingh]

fyi @pshipton Identical to ibmruntimes/openj9-openjdk-jdk16#29. For supporting OJDK MHs.

[pshipton]

ibmruntimes/openj9-openjdk-jdk16@f6e2b25af55d44 also added a test MethodHandlesCollectArgsTest.java, I think we want that as well?

[pshipton]

Maybe we'd have to exclude the test until we switch?

[pshipton]

We're backporting an OpenJDK change, so we should backport the Oracle copyright, not add an IBM copyright.

[pshipton]

Why do we need to backport this if OpenJDK hasn't bothered to backport to 8, 11?

[babsingh]

Why do we need to backport this if OpenJDK hasn't bothered to backport to 8, 11?

Because there are tests in OpenJ9, which will fail without this fix, once OJDK MHs are enabled: eclipse-openj9/openj9#11922.

We're backporting an OpenJDK change, so we should backport the Oracle copyright, not add an IBM copyright.

Updated Oracle copyright and removed IBM copyright.

ibmruntimes/openj9-openjdk-jdk16@f6e2b25af55d44 also added a test MethodHandlesCollectArgsTest.java, I think we want that as well?

This test won't work with JDK8 because MethodHandles.empty is not available in JDK8. We can either update the test or rely on the existing OJ9 tests for coverage in JDK8. For JDK11, I have included the test in ibmruntimes/openj9-openjdk-jdk11#502.

Maybe we'd have to exclude the test until we switch?

We do not need to exclude the test because the test should correctly work with OJ9 MHs

[pshipton]

jenkins test sanity,extended alinux jdk8

[pshipton]

Oops, PR testing isn't useful here since the class being modified isn't used atm. Luckily it failed anyway because I got the platform name wrong.