Change method lookup infrastructure to use the trait methods. Instead

of tracking individual candidates per impl, we just track one
candidate for the extension trait itself, and let the trait resolution
handle walking the individual impls and so forth. Also change the
interface to report back a richer notion of error.

src/librustc/middle/ty.rs

@@ -120,6 +120,13 @@ impl ImplOrTraitItem {
             TypeTraitItem(ref associated_type) => associated_type.container,
         }
     }
+
+    pub fn as_opt_method(&self) -> Option<Rc<Method>> {
+        match *self {
+            MethodTraitItem(ref m) => Some((*m).clone()),
+            TypeTraitItem(_) => None
+        }
+    }
 }
 
 #[deriving(Clone)]
@@ -1240,6 +1247,10 @@ impl Generics {
 }
 
 impl TraitRef {
+    pub fn new(def_id: ast::DefId, substs: Substs) -> TraitRef {
+        TraitRef { def_id: def_id, substs: substs }
+    }
+
     pub fn self_ty(&self) -> ty::t {
         self.substs.self_ty().unwrap()
     }