John Rose a écrit :
> On Oct 9, 2008, at 1:49 PM, Rémi Forax wrote:
>
>> With this encoding, you have a receiveless 5-byte opcode.
>>
>
> Good point. But, someone is sure to complain that the normal uses of
> the opcode do not allow null. Then we get around to those other
> designs we talked about at the Summit, where the metaobject is passed
> as the receiver. This is clever, but it optimizes the slow path
> instead of the fast path, so I don't think that's what we want.
not if we mandate ACONST_NULL, in that case, it can be removed from the
fast path.
>
> -- John
Rémi
_______________________________________________
mlvm-dev mailing list
[email protected]
http://mail.openjdk.java.net/mailman/listinfo/mlvm-dev
