At 01:41 PM 1/12/2003 -0800, Jerry Schwarz wrote:
>I think the correct thing to say is that Object.notify and
>Thread.interrupt respect the "happens before" relationship. Because
>"happens before" totally orders events in a given thread this implies the
>above form.
I believe that this is the effect of the "FOR A NOTIFICATION" addition to
the specification of Thread.interrupt() anyway, because of the way wait()
describes actions on the otherwise invisible wait set. The issue was that
there remained questions about the possibility of reordering operations,
but my impression is that that issue is now resolved to the effect that
such reordering is not permitted.
So all that's needed now is a general agreement that this is the intended
consequence of the change to the spec, and that it can be implemented in an
acceptably efficient manner.
Any takers?
Sylvia.
-------------------------------
JavaMemoryModel mailing list - http://www.cs.umd.edu/~pugh/java/memoryModel
This archive was generated by hypermail 2b29 : Thu Oct 13 2005 - 07:00:55 EDT