Proposal: java.lang.reflect.Proxy and default methods

Mandy Chung mandy.chung at oracle.com
Mon Jun 6 20:58:37 UTC 2016


Hi Peter,

Thanks for the proposal. This feature has been lacking.  When this subject was brought, I also have similiar thought to provide a way in Proxy class for InvocationHandler to invoke a default method (but of course no time to put into it).

I appreciate your contribution and good work.  I support to add this feature in a future release.

I personally don’t feel comfortable to absorb this small API targetting for JDK 9 (since I don’t have the cycle to shepherd this in the next few months). I may be overly conversative but I won’t take security assessment lightly.  Maybe someone else is able to help you move this forward before I am available.

Mandy

> On Jun 3, 2016, at 7:58 AM, Peter Levart <peter.levart at gmail.com> wrote:
> 
> Hi,
> 
> Since Java SE 8 introduced default methods in interfaces there was a question what to do with java.lang.reflect.Proxy API. Nothing was done to the API at that time, so the default behavior is to proxy default methods too. InvocationHandler gets invoked for default methods, but it has not provision to forward such calls to the default implementations in the interfaces.
> 
> I propose a simple API addition that allows calling super default methods in proxy instances:
> 
> http://cr.openjdk.java.net/~plevart/jdk9-dev/Proxy.invokeSuperDefaults/webrev.02/
> 
> With this addition one can simply decide in the InvocationHandler what to do with invocations to default methods and can forward such invocation to the default implementation:
> 
> public class Test {
> 
>    interface I1 {
>        default void m() {
>            System.out.println("  default I1.m() called");
>        }
>    }
> 
>    interface I2 {
>        default void m() {
>            System.out.println("  default I2.m() called");
>        }
>    }
> 
>    interface I12 extends I1, I2 {
>        @Override
>        void m();
> 
>        default int sum(int a, int b) {
>            return a + b;
>        }
> 
>        default Object[] concat(Object first, Object... rest) {
>            Object[] result = new Object[1 + rest.length];
>            result[0] = first;
>            System.arraycopy(rest, 0, result, 1, rest.length);
>            return result;
>        }
>    }
> 
>    public static void main(String[] args) {
> 
>        InvocationHandler h = (proxy, method, params) -> {
>            System.out.println("\nInvocationHandler called for: " + method +
>                               " with parameters: " + Arrays.toString(params));
>            if (method.isDefault()) {
>                try {
>                    return Proxy.invokeSuper(proxy, method, params);
>                } catch (InvocationTargetException e) {
>                    throw e.getCause();
>                }
>            } else {
>                switch (method.getName()) {
>                    case "m":
>                        System.out.println("  abstract I12.m(): called");
>                        return null;
>                    default:
>                        throw new UnsupportedOperationException(
>                            "Unsupported method: " + method);
>                }
>            }
>        };
> 
>        I1 i1 = (I1) Proxy.newProxyInstance(
>            I1.class.getClassLoader(), new Class<?>[]{I1.class}, h);
>        i1.m();
> 
>        I2 i2 = (I2) Proxy.newProxyInstance(
>            I2.class.getClassLoader(), new Class<?>[]{I2.class}, h);
>        i2.m();
> 
>        I12 i12 = (I12) Proxy.newProxyInstance(
>            I12.class.getClassLoader(), new Class<?>[]{I12.class}, h);
>        i12.m();
> 
>        System.out.println("  1 + 2 = " + i12.sum(1, 2));
>        System.out.println("  [1] concat [2, 3, 4] = " +
>                           Arrays.toString(i12.concat(1, 2, 3, 4)));
>    }
> }
> 
> 
> I know FC date is over, but this is really a small change and I have heard several people that such feature is missing from the Proxy API.
> 
> I'm prepared to create jtreg tests covering the specification if this proposal is accepted.
> 
> Regards, Peter
> 
> 
> 



More information about the core-libs-dev mailing list