[ Doug Lea ] [concurrency-interest] Re: AtomicInteger and AtomicLong should implement Number

Doug Lea dl@cs.oswego.edu
Thu, 15 Jan 2004 00:31:11 -0500

Gregg G. Wonderly wrote:
>>3. In those cases where people want value-based equality etc
>>they can either rely on autoboxing, which provides reliable
>>semantics at the possible cost of overhead, or create subclasses
>>that add on particular equals, hashCode, and/or compareTo methods.
> AtomicX classes are final last I looked

No; they are nonfinal but all atomic methods are final. (Although the
early-access version had some inconsitencies across classes in how this
was done (sorry.))

And we remain confident that the current AtomicX APIs are both
conceptually and pragmatically about right.