My application initializes an object graph in $rootScope, like this ...
var myApp = angular.module(\'myApp\', []);
myApp.run(function ($rootScope) {
$ro
If you want to update root scope's object, inject $rootScope
into your controller:
myApp.controller('MyController', function ($scope, $timeout, $rootScope) {
$rootScope.myObject = { value: 3 };
$timeout(function() {
$rootScope.myObject = { value: 4 };
$timeout(function () {
$rootScope.myObject = { value: 5 };
}, 1000);
}, 1000);
});
Demo fiddle
After you update the $rootScope call $rootScope.$apply() to update the bindings.
Think of modifying the scopes as an atomic operation and $apply() commits those changes.
Except for very, very rare cases or debugging purposes, doing this is just BAD practice (or an indication of BAD application design)!
For the very, very rare cases (or debugging), you can do it like this:
$rootScope
by accessing .scope().$root
. (There are other ways as well.)$rootScope.$apply()
, so Angular will know something is going on and do its magic.E.g.:
function badPractice() {
var $body = angular.element(document.body); // 1
var $rootScope = $body.scope().$root; // 2
$rootScope.$apply(function () { // 3
$rootScope.someText = 'This is BAD practice :(';
});
}
See, also, this short demo.
EDIT
Angular 1.3.x introduced an option to disable debug-info from being attached to DOM elements (including the scope
): $compileProvider.debugInfoEnabled()
It is advisable to disable debug-info in production (for performance's sake), which means that the above method would not work any more.
If you just want to debug a live (production) instance, you can call angular.reloadWithDebugInfo(), which will reload the page with debug-info enabled.
Alternatively, you can go with Plan B (accessing the $rootScope
through an element's injector):
function badPracticePlanB() {
var $body = angular.element(document.body); // 1
var $rootScope = $body.injector().get('$rootScope'); // 2b
$rootScope.$apply(function () { // 3
$rootScope.someText = 'This is BAD practice too :(';
});
}