Document the edge case in a FIXME as caused by define_method use

* We should rip out the fix when we update the define_method
  implementation to convert those blocks to real method scopes.

Author: subbuss

core/src/main/java/org/jruby/ir/runtime/IRRuntimeHelpers.java

@@ -742,6 +742,28 @@ public static RubyModule findInstanceMethodContainer(ThreadContext context, Dyna
                 // - in the rare case where it is not (looks like it can
                 //   happen in some testing frameworks), we have to add
                 //   the method to self itself => its metaclass.
+                //
+                // SSS FIXME: Looks like this rare case happens when
+                // the closure is used in a "define_method &block" scenario
+                // => in reality the scope is not a closure but an
+                // instance_method. So, when we fix define_method implementation
+                // to actually convert blocks to real instance_method scopes,
+                // we will not have this edge case since the code will then
+                // be covered by the (scopeType == IRScopeType.INSTANCE_METHOD)
+                // scenario below. Whenever we get to fixing define_method
+                // implementation, we should rip out this code here.
+                //
+                // Verify that this test runs:
+                // -------------
+                //   require "minitest/autorun"
+                //
+                //   describe "A" do
+                //     it "should do something" do
+                //       def foo
+                //       end
+                //     end
+                //   end
+                // -------------
                 case MODULE_EVAL  : return self instanceof RubyModule ? (RubyModule) self : self.getMetaClass();
                 case INSTANCE_EVAL: return self.getSingletonClass();
                 case BINDING_EVAL : ds = ds.getParentScope(); break;