On 06/03/2011 11:19 AM, Emmanuel Castro wrote:
> I noticed that the same kind of strange properties exists from 
> the beginning in the Java API on the Thread class. When the 
> Thread.isAlive() returns true, you're not sure that it is still true 
> at the next instruction.
> If one follows the same naming scheme, it would yield to Thread.hasDied().
>
> Emmanuel CASTRO

Yes.

The other problem is that SwitchPoint.gWT can be made fast but
it's harder to get the same speed if someone uses 
SwitchPoint.hasBeenInvalidated
in a fast path.
We really don't want bug reports saying that hasBeenInvalidated is dead slow
forcing VM implementers to find a way to optimize hasBeenInvalidated
knowing that SwitchPoint.gWT is the way to go.

Rémi

_______________________________________________
mlvm-dev mailing list
[email protected]
http://mail.openjdk.java.net/mailman/listinfo/mlvm-dev

Reply via email to