Interface StateDeclaration

An abstract state can never be directly activated. Use an abstract state to provide inherited properties (url, resolve, data, etc) to children states.

This is a spot for you to store inherited state metadata. Child states' data object will prototypally inherit from their parent state.

This is a good spot to put metadata such as requiresAuth.

Note: because prototypal inheritance is used, changes to parent data objects reflect in the child data objects. Care should be taken if you are using hasOwnProperty on the data object. Properties from parent objects will return false for hasOwnProperty.

All parameters on the state will use this value for dynamic as a default. Individual parameters may override this default using ParamDeclaration.dynamic in the params block.

Note: this value overrides the dynamic value on a custom parameter type (ParamTypeDefinition.dynamic).

The lazyLoad function is invoked before the state is activated. The transition waits while the code is loading.

The function should load the code that is required to activate the state. For example, it may load a component class, or some service code. The function must return a promise which resolves when loading is complete.

For example, this code lazy loads a service before the abc state is activated:

.state('abc', {
  lazyLoad: (transition, state) => import('./abcService')
}

The abcService file is imported and loaded (it is assumed that the abcService file knows how to register itself as a service).

Lifecycle

• The lazyLoad function is invoked if a transition is going to enter the state.

• The function is invoked before the transition starts (using an onBefore transition hook).

• The function is only invoked once; while the lazyLoad function is loading code, it will not be invoked again. For example, if the user double clicks a ui-sref, lazyLoad is only invoked once even though there were two transition attempts. Instead, the existing lazy load promise is re-used.

• When the promise resolves successfully, the lazyLoad property is deleted from the state declaration.

• If the promise resolves to a LazyLoadResult which has an array of states, those states are registered.

• The original transition is retried (this time without the lazyLoad property present).

• If the lazyLoad function fails, then the transition also fails. The failed transition (and the lazyLoad function) could potentially be retried by the user.

Lazy loading state definitions (Future States)

State definitions can also be lazy loaded. This might be desirable when building large, multi-module applications.

To lazy load state definitions, a Future State should be registered as a placeholder. When the state definitions are lazy loaded, the Future State is deregistered.

A future state can act as a placeholder for a single state, or for an entire module of states and substates. A future state should have:

• A name which ends in .**. A future state's name property acts as a wildcard Glob. It matches any state name that starts with the name (including child states that are not yet loaded).
• A url prefix. A future state's url property acts as a wildcard. UI-Router matches all paths that begin with the url. It effectively appends .* to the internal regular expression. When the prefix matches, the future state will begin loading.
• A lazyLoad function. This function should should return a Promise to lazy load the code for one or more StateDeclaration objects. It should return a LazyLoadResult. Generally, one of the lazy loaded states should have the same name as the future state. The new state will then replace the future state placeholder in the registry.

Additional resources

For in depth information on lazy loading and Future States, see the Lazy Loading Guide.

Example: states.js

// This child state is a lazy loaded future state
// The `lazyLoad` function loads the final state definition
{
  name: 'parent.**',
  url: '/parent',
  lazyLoad: () => import('./lazy.states.js')
}

Example: lazy.states.js

This file is lazy loaded. It exports an array of states.

import {ChildComponent} from "./child.component.js";
import {ParentComponent} from "./parent.component.js";

// This fully defined state replaces the future state
let parentState = {
  // the name should match the future state
  name: 'parent',
  url: '/parent/:parentId',
  component: ParentComponent,
  resolve: {
    parentData: ($transition$, ParentService) =>
        ParentService.get($transition$.params().parentId)
  }
}

let childState = {
  name: 'parent.child',
  url: '/child/:childId',
  params: {
    childId: "default"
  },
  resolve: {
    childData: ($transition$, ChildService) =>
        ChildService.get($transition$.params().childId)
  }
};

// This array of states will be registered by the lazyLoad hook
let lazyLoadResults = {
  states: [ parentState, childState ]
};

export default lazyLoadResults;

param

the Transition that is activating the future state

param

the StateDeclaration that the lazyLoad function is declared on

returns

a Promise to load the states. Optionally, if the promise resolves to a LazyLoadResult, the states will be registered with the StateRegistry.

A unique state name, e.g. "home", "about", "contacts". To create a parent/child state use a dot, e.g. "about.sales", "home.newest".

Note: [State] objects require unique names. The name is used like an id.

Example:

.state({
  name: 'mystate',
  onEnter: function(trans, state) {
    console.log("Entering " + state.name);
  }
});

Note: The above onEnter on the state declaration is effectively sugar for:

transitionService.onEnter({ entering: 'mystate' }, function(trans, state) {
  console.log("Entering " + state.name);
});

Example:

.state({
  name: 'mystate',
  onExit: function(trans, state) {
    console.log("Leaving " + state.name);
  }
});

Note: The above onRetain on the state declaration is effectively sugar for:

transitionService.onExit({ exiting: 'mystate' }, function(trans, state) {
  console.log("Leaving " + state.name);
});

Example:

.state({
  name: 'mystate',
  onRetain: function(trans, state) {
    console.log(state.name + " is still active!");
  }
});

Note: The above onRetain on the state declaration is effectively sugar for:

transitionService.onRetain({ retained: 'mystate' }, function(trans, state) {
  console.log(state.name + " is still active!");
});

An object which optionally configures parameters declared in the url, or defines additional non-url parameters. For each parameter being configured, add a ParamDeclaration keyed to the name of the parameter.

Example:

params: {
  param1: {
   type: "int",
   array: true,
   value: []
  },
  param2: {
    value: "index"
  }
}

Normally, a state's parent is implied from the state's name, e.g., "parentstate.childstate".

Alternatively, you can explicitly set the parent state using this property. This allows shorter state names, e.g., <a ui-sref="childstate">Child</a> instead of `Child

When using this property, the state's name should not have any dots in it.

Example:

var parentstate = {
  name: 'parentstate'
}
var childstate = {
  name: 'childstate',
  parent: 'parentstate'
  // or use a JS var which is the parent StateDeclaration, i.e.:
  // parent: parentstate
}

If this property is defined, a Transition directly to this state will be redirected based on the property's value.

• If the value is a string, the Transition is redirected to the state named by the string.

• If the property is an object with a state and/or params property, the Transition is redirected to the named state and/or params.

• If the value is a TargetState the Transition is redirected to the TargetState

• If the property is a function:

  • The function is called with the current Transition
  • The return value is processed using the previously mentioned rules.
  • If the return value is a promise, the promise is waited for, then the resolved async value is processed using the same rules.

Note: redirectTo is processed as an onStart hook, before LAZY resolves. If your redirect function relies on resolve data, get the Transition.injector and get a promise for the resolve data using UIInjector.getAsync.

Example:

// a string
.state('A', {
  redirectTo: 'A.B'
})

// a {state, params} object
.state('C', {
  redirectTo: { state: 'C.D', params: { foo: 'index' } }
})

// a fn
.state('E', {
  redirectTo: () => "A"
})

// a fn conditionally returning a {state, params}
.state('F', {
  redirectTo: (trans) => {
    if (trans.params().foo < 10)
      return { state: 'F', params: { foo: 10 } };
  }
})

// a fn returning a promise for a redirect
.state('G', {
  redirectTo: (trans) => {
    let svc = trans.injector().get('SomeAsyncService')
    let promise = svc.getAsyncRedirectTo(trans.params.foo);
    return promise;
  }
})

// a fn that fetches resolve data
.state('G', {
  redirectTo: (trans) => {
    // getAsync tells the resolve to load
    let resolvePromise = trans.injector().getAsync('SomeResolve')
    return resolvePromise.then(resolveData => resolveData === 'login' ? 'login' : null);
  }
})

deprecated

use either dynamic or ParamDeclaration.dynamic

The resolve: property defines data (or other dependencies) to be fetched asynchronously when the state is being entered. After the data is fetched, it may be used in views, transition hooks or other resolves that belong to this state. The data may also be used in any views or resolves that belong to nested states.

As an array

Each array element should be a ResolvableLiteral object.

Example:

The user resolve injects the current Transition and the UserService (using its token, which is a string). The [[ResolvableLiteral.resolvePolicy]] sets how the resolve is processed. The user data, fetched asynchronously, can then be used in a view.

var state = {
  name: 'user',
  url: '/user/:userId
  resolve: [
    {
      token: 'user',
      policy: { when: 'EAGER' },
      deps: ['UserService', Transition],
      resolveFn: (userSvc, trans) => userSvc.fetchUser(trans.params().userId) },
    }
  ]
}

Note: an Angular 2 style useFactory provider literal may also be used. See ProviderLike.

Example:

resolve: [
  { provide: 'token', useFactory: (http) => http.get('/'), deps: [ Http ] },
]

As an object

The resolve property may be an object where:

• Each key (string) is the name of the dependency.
• Each value (function) is an injectable function which returns the dependency, or a promise for the dependency.

This style is based on AngularJS injectable functions, but can be used with any UI-Router implementation. If your code will be minified, the function should be "annotated" in the AngularJS manner.

AngularJS Example:

resolve: {
  // If you inject `myStateDependency` into a controller, you'll get "abc"
  myStateDependency: function() {
    return "abc";
  },
  // Dependencies are annotated in "Inline Array Annotation"
  myAsyncData: ['$http', '$transition$' function($http, $transition$) {
    // Return a promise (async) for the data
    return $http.get("/foos/" + $transition$.params().foo);
  }]
}

Note: You cannot specify a policy for each Resolvable, nor can you use non-string tokens when using the object style resolve: block.

Lifecycle

Since a resolve function can return a promise, the router will delay entering the state until the promises are ready. If any of the promises are rejected, the Transition is aborted with an Error.

By default, resolves for a state are fetched just before that state is entered. Note that only states which are being entered during the Transition have their resolves fetched. States that are "retained" do not have their resolves re-fetched.

If you are currently in a parent state parent and are transitioning to a child state parent.child, the previously resolved data for state parent can be injected into parent.child without delay.

Any resolved data for parent.child is retained until parent.child is exited, e.g., by transitioning back to the parent state.

Because of this scoping and lifecycle, resolves are a great place to fetch your application's primary data.

Injecting resolves into other things

During a transition, Resolve data can be injected into:

• Views (the components which fill a ui-view tag)
• Transition Hooks
• Other resolves (a resolve may depend on asynchronous data from a different resolve)

Injecting other things into resolves

Resolve functions usually have dependencies on some other API(s). The dependencies are usually declared and injected into the resolve function. A common pattern is to inject a custom service such as UserService. The resolve then delegates to a service method, such as UserService.list();

Special injectable tokens

• UIRouter: The UIRouter instance which has references to all the UI-Router services.
• Transition: The current Transition object; information and API about the current transition, such as "to" and "from" State Parameters and transition options.
• '$transition$': A string alias for the Transition injectable
• '$state$': For onEnter/onExit/onRetain, the state being entered/exited/retained.
• Other resolve tokens: A resolve can depend on another resolve, either from the same state, or from any parent state.

Example:

// Injecting a resolve into another resolve
resolve: [
  // Define a resolve 'allusers' which delegates to the UserService.list()
  // which returns a promise (async) for all the users
  { provide: 'allusers', useFactory: (UserService) => UserService.list(), deps: [UserService] },

  // Define a resolve 'user' which depends on the allusers resolve.
  // This resolve function is not called until 'allusers' is ready.
  { provide: 'user', (allusers, trans) => _.find(allusers, trans.params().userId, deps: ['allusers', Transition] }
}

This should be an ResolvePolicy object.

It can contain the following optional keys/values:

• when: (optional) defines when the resolve is fetched. Accepted values: "LAZY" or "EAGER"
• async: (optional) if the transition waits for the resolve. Accepted values: "WAIT", "NOWAIT", CustomAsyncPolicy

See ResolvePolicy for more details.

A URL fragment (with optional parameters) which is used to match the browser location with this state.

This fragment will be appended to the parent state's URL in order to build up the overall URL for this state. See UrlMatcher for details on acceptable patterns.

examples

url: "/home"
// Define a parameter named 'userid'
url: "/users/:userid"
// param 'bookid' has a custom regexp
url: "/books/{bookid:[a-zA-Z_-]}"
// param 'categoryid' is of type 'int'
url: "/books/{categoryid:int}"
// two parameters for this state
url: "/books/{publishername:string}/{categoryid:int}"
// Query parameters
url: "/messages?before&after"
// Query parameters of type 'date'
url: "/messages?{before:date}&{after:date}"
// Path and query parameters
url: "/messages/:mailboxid?{before:date}&{after:date}"

An optional object which defines multiple views, or explicitly targets specific named ui-views.

• What is a view config

• What is a ui-view

• Shorthand controller/template

• Incompatible with ^

  Examples:

  Targets three named ui-views in the parent state's template

Example:

views: {
  header: {
    controller: "headerCtrl",
    templateUrl: "header.html"
  }, body: {
    controller: "bodyCtrl",
    templateUrl: "body.html"
  }, footer: {
    controller: "footCtrl",
    templateUrl: "footer.html"
  }
}

example

// Targets named ui-view="header" from ancestor state 'top''s template, and
// named `ui-view="body" from parent state's template.
views: {
  'header@top': {
    controller: "msgHeaderCtrl",
    templateUrl: "msgHeader.html"
  }, 'body': {
    controller: "messagesCtrl",
    templateUrl: "messages.html"
  }
}