Method references in annotations?
Jason Schroeder
shrode at subnature.com
Mon Aug 15 15:56:39 PDT 2011
Neal,
Since I now read that there are internal conversations, any
implementation experiment
I have here is noise, until those email threads appear.
To answer your question:
The signature of Method.invoke implements all methods. And therefore
the signature of Method.invoke is the uninteresting, unused SAM for
all methods in Java. So for the purpose of boxing a method reference,
and not adding a plain literal hack of "Method m = #A.foo," it is
possible to explore the ramifications of making a SAM interface for
Method.invoke and expecting an optimization/special-case of:
MethodInvoke m = #A.foo
to play by the rules of SAM but return the actual Method. As a
starting point, then one can begin to elide the needs for some
similarity across method references by specifying a boxing-like rule.
My aim was to reach into MethodHandles and get the annotation
requirement for free.
Jason
On Sun, Aug 14, 2011 at 2:46 PM, Neal Gafter <neal at gafter.com> wrote:
> I am not sure what you're trying to say in this message. Method.invoke is a
> method. SAMs are interface types. I don't know what it means to treat a
> method as a type. Do you mean to suggest treating java.lang.Method as a
> SAM? It has little in common with them.
More information about the lambda-dev
mailing list