On 10/18/2012 6:11 PM, David Holmes wrote:
Hi Martin,
On 19/10/2012 10:58 AM, Martin Buchholz wrote:
http://code.google.com/p/guava-libraries/source/browse/guava-testlib/src/com/google/common/testing/GcFinalization.java
That code uses runFinalization in places which means it is not
exercising the primary finalization mechanism. But looking further
into the whole thing I don't think it really makes much difference
semantically afterall. All that changes is which thread executes the
finalizer - the object comes from the same ReferenceQueue regardless.
Which means that runFinalization() will really only help if the main
finalizer thread is delayed executing another finalizer - which for
some tests may be caused by compilation of the finalize method.
So Chris/Mandy: I think the original suggested fix is fine here.
Thanks David. I looked at the Finalizer class and also concur that what
Chris has is fine.
Chris - do you have the committer right to push to jdk8/tl? If not, I
can push it for you.
Mandy
