Minor fix to PyJavaFunction 2-arg call

Drafting error fixed. Some improvements to comments.

Author: jeff5

rt4core/src/main/java/uk/co/farowl/vsj4/runtime/PyJavaFunction.java

@@ -217,7 +217,10 @@ static PyJavaFunction forStaticMethod(ArgParser ap,
      * Construct a {@code PyJavaFunction} from an {@link ArgParser} and
      * {@code MethodHandle} for a {@code __new__} method. Although
      * {@code __new__} is a static method the {@code PyJavaFunction} we
-     * produce is bound to the defining {@code PyType self}.
+     * produce is bound (by setting {@link #self}) to the defining
+     * {@code PyType self}. The {@link #handle} derived from
+     * {@code method} is <i>not</i> bound to {@code self}, but
+     * <i>does</i> add a filter to check the type of the first argument.
      *
      * @param ap argument parser (provides argument names etc.)
      * @param method raw handle to the method defined
@@ -662,9 +665,8 @@ public Object call(Object a0) throws Throwable {
         }
 
         @Override
-        public Object call(Object self, Object a0, Object a1)
-                throws Throwable {
-            return handle.invokeExact(self, a0, a1);
+        public Object call(Object a0, Object a1) throws Throwable {
+            return handle.invokeExact(a0, a1);
         }
     }
 

rt4core/src/main/java/uk/co/farowl/vsj4/runtime/subclass/SubclassSpec.java

@@ -123,6 +123,13 @@ protected final SubclassSpec freeze() {
             }
 
             // FIXME : (maybe order and) freeze interfaces.
+            /*
+             * Order of interfaces is *not* significant in Java and
+             * treating them as ordered in the MRO is perhaps behind
+             * certain problems in Jython 2 such as #70 and #391. Yet we
+             * must (I think) acknowledge them when looking for a
+             * representation, and probably as bases in Python.
+             */
 
             /*
              * The specification calls for instances to have a __dict__