mocking with amd

gistfile1.js

// ----------------------------------------
// /path/to/dependency.js
define(function() {
  return {
    doSomethingWithIt: function() {
      // blah
    }
  };
});

// ----------------------------------------
// my_object.js
define(['/path/to/dependency'], function(myDependency) {

  function MyObject() {
    this.dependency = myDependency;
  }
  MyObject.prototype = {
    one: function() {
      this.dependency.doSomethingWithIt();
      return 'blah';
    }
  };

  return MyObject;
});

// ----------------------------------------
// my_object_spec.js
// the object under test is my_object, so ideally
// I'd like to mock/stub myDependency
define(['my_object', '/path/to/dependency'], function(MyObject, myDependency) {

  beforeEach(function() {
    sinon.stub(myDependency, 'doSomethingWithIt').returns('whatever');
    this.subject = new MyObject();
  });

  afterEach(function() {
    myDependency.doSomethingWithIt.restore();
  });

  it('should call one() as we expect', function() {
    expect(this.subject.one(), 'blah');
  });
});

I don't think you'd need to do this. Creating a property to hold a dependency is not needed if you already have the dependency available in closure scope. So in your one() method, you just run myDependency.doSomethingWithIt() directly. It's a bit hard to see what you're getting at without seeing a real-world example.

See you at Build. I will be found nursing a hangover in the Malmaison.

Hey @froots, just ran into a similar problem again this morning and I have a more concrete example that may explain things a little better of what we're trying to achieve: https://gist.github.com/3187758