When you declare ng-app here:
<div ng-app='app' ng-controller='controller1'>
</div>
Angular will look for a module definition with the name app. Something like this:
angular.module('app', []); // notice the []
Inside the second parameter [] array, Angular wants the dependent modules. So, in order to include controller1 from foo module, you would do this:
angular.module('app', ['foo']);
And from the bar module:
angular.module('app', ['bar']);
In each of these, there is only a single controller1 named controller.
So, what happens when you register both the foo and bar modules? I would think that the last one wins. So, if you define app to be:
angular.module('app', ['foo', 'bar']);
Then, the bar module's controller1 will overwrite the name controller1 inside the injector.
So, there is no built-in mechanism that allows for the same name to be applied across modules. Because of this, you can employ naming schemes to make sure that the controller is unique:
angular.module('bar').controller('bar.controller1', ['$scope', function($scope) {
});
How does angular know that 'controller1' resides in module 'foo'?
At the time that Angular is resolving controller1, it doesn't know what module it is in. Once the module is a dependent (on the module defined in ng-app), all of the controllers (in all modules) are only resolved off of their names.
