RFR 7118066: Warnings in java.util.concurrent package (original) (raw)

Chris Hegarty chris.hegarty at oracle.com
Thu Dec 8 15:30:18 UTC 2011

On 08/12/2011 15:21, Mike Duigou wrote:

Was the SnychronousQueue test recently added or did something else change to cause the failure?

This failure has not happened yet, but is will when I push the "warning" changes for j.u.c ( with some minor impl changes ).

SynchronousQueue.iterator() no longer returns Collections.emptyIterator(), it will return its own private EmptyIterator implementation. Collections.emptyIterator was added in 1.7 and we need to able to compile with 1.6. There is no spec requirement on SynchronousQueue.iterator(). I think the test just took advantage the private implementation of SynchronousQueue.

-Chris.

Mike On Dec 8 2011, at 03:33 , Chris Hegarty wrote:

On 08/12/2011 05:13, David Holmes wrote: ....

I also think the SynchronousQueue has no business being used in this test. Though I confess I'm unclear what this test is for - is it for emptyIterator() or for "iterators of empty collections" ? Thanks David, so the proposed change is to simple remove SynchronousQueue. --- a/test/java/util/Collections/EmptyIterator.java Wed Dec 07 10:17:34 2011 +0000 +++ b/test/java/util/Collections/EmptyIterator.java Thu Dec 08 11:28:16 2011 +0000 @@ -35,8 +35,6 @@ public class EmptyIterator { void test(String[] args) throws Throwable { testEmptyCollection(Collections.emptyList()); testEmptyCollection(Collections.emptySet()); - testEmptyCollection(new java.util.concurrent. - SynchronousQueue()); testEmptyMap(Collections.<Object, Object>emptyMap()); Since it is trivial, I'll push it with the warning changes. -Chris.

David -----

More information about the core-libs-dev mailing list