Service Workers

1. Motivations

This section is non-normative.

Web Applications traditionally assume that the network is reachable. This assumption pervades the platform. HTML documents are loaded over HTTP and traditionally fetch all of their sub-resources via subsequent HTTP requests. This places web content at a disadvantage versus other technology stacks.

The service worker is designed first to redress this balance by providing a Web Worker context, which can be started by a runtime when navigations are about to occur. This event-driven worker is registered against an origin and a path (or pattern), meaning it can be consulted when navigations occur to that location. Events that correspond to network requests are dispatched to the worker and the responses generated by the worker may override default network stack behavior. This puts the service worker, conceptually, between the network and a document renderer, allowing the service worker to provide content for documents, even while offline.

Web developers familiar with previous attempts to solve the offline problem have reported a deficit of flexibility in those solutions. As a result, the service worker is highly procedural, providing a maximum of flexibility at the price of additional complexity for developers. Part of this complexity arises from the need to keep service workers responsive in the face of a single-threaded execution model. As a result, APIs exposed by service workers are almost entirely asynchronous, a pattern familiar in other JavaScript contexts but accentuated here by the need to avoid blocking document and resource loading.

Developers using the HTML5 Application Cache have also reported that several attributes of the design contribute to unrecoverable errors. A key design principle of the service worker is that errors should always be recoverable. Many details of the update process of service workers are designed to avoid these hazards.

Service workers are started and kept alive by their relationship to events, not documents. This design borrows heavily from developer and vendor experience with Shared Workers and Chrome Background Pages. A key lesson from these systems is the necessity to time-limit the execution of background processing contexts, both to conserve resources and to ensure that background context loss and restart is top-of-mind for developers. As a result, service workers bear more than a passing resemblance to Chrome Event Pages, the successor to Background Pages. Service workers may be started by user agents without an attached document and may be killed by the user agent at nearly any time. Conceptually, service workers can be thought of as Shared Workers that can start, process events, and die without ever handling messages from documents. Developers are advised to keep in mind that service workers may be started and killed many times a second.

Service workers are generic, event-driven, time-limited script contexts that run at an origin. These properties make them natural endpoints for a range of runtime services that may outlive the context of a particular document, e.g. handling push notifications, background data synchronization, responding to resource requests from other origins, or receiving centralized updates to expensive-to-calculate data (e.g., geolocation or gyroscope).

2. Model

2.1. Service Worker

A service worker is a type of web worker. A service worker executes in the registering service worker client's origin.

A service worker has an associated state, which is one of "parsed", "installing", "installed", "activating", "activated", and "redundant". It is initially "parsed".

A service worker has an associated script url (a URL).

A service worker has an associated type which is either "classic" or "module". Unless stated otherwise, it is "classic".

A service worker has an associated containing service worker registration (a service worker registration), which contains itself.

A service worker has an associated global object (a ServiceWorkerGlobalScope object or null).

A service worker has an associated script resource (a script), which represents its own script resource. It is initially set to null.

A script resource has an associated has ever been evaluated flag. It is initially unset.

A script resource has an associated policy container (a policy container). It is initially a new policy container.

A service worker has an associated script resource map which is an ordered map where the keys are URLs and the values are responses.

A service worker has an associated set of used scripts (a set) whose item is a URL. It is initially a new set.

Note: The set of used scripts is only used to prune unused resources from a new worker’s map after installation, that were populated based on the old worker’s map during the update check.

A service worker has an associated skip waiting flag. Unless stated otherwise it is unset.

A service worker has an associated classic scripts imported flag. It is initially unset.

A service worker has an associated set of event types to handle (a set) whose item is an event listener’s event type. It is initially a new set.

A service worker has an associated set of extended events (a set) whose item is an ExtendableEvent. It is initially a new set.

A service worker has an associated start status which can be null or a Completion. It is initially null.

A service worker is said to be running if its event loop is running.

2.1.1. Lifetime

The lifetime of a service worker is tied to the execution lifetime of events and not references held by service worker clients to the ServiceWorker object.

A user agent may terminate service workers at any time it:

Has no event to handle.
Detects abnormal operation: such as infinite loops and tasks exceeding imposed time limits (if any) while handling the events.

2.1.2. Events

The Service Workers specification defines service worker events (each of which is an event) that include (see the list):

Lifecycle events: install and activate.
Functional events: fetch and the events defined by other specifications that extend the Service Workers specification. (See the list.)
message and messageerror.

2.2. Service Worker Timing

Service workers mark certain points in time that are later exposed by the navigation timing API.

A service worker timing info is a struct. It has the following items:

start time: A DOMHighResTimeStamp, initially 0.
fetch event dispatch time: A DOMHighResTimeStamp, initially 0.

2.3. Service Worker Registration

A service worker registration is a tuple of a scope url, a storage key, and a set of service workers, an installing worker, a waiting worker, and an active worker. A user agent may enable many service worker registrations at a single origin so long as the scope url of the service worker registration differs. A service worker registration of an identical scope url when one already exists in the user agent causes the existing service worker registration to be replaced.

A service worker registration has an associated storage key (a storage key).

A service worker registration has an associated scope url (a URL).

A service worker registration has an associated installing worker (a service worker or null) whose state is "installing". It is initially set to null.

A service worker registration has an associated waiting worker (a service worker or null) whose state is "installed". It is initially set to null.

A service worker registration has an associated active worker (a service worker or null) whose state is either "activating" or "activated". It is initially set to null.

A service worker registration has an associated last update check time. It is initially set to null.

A service worker registration is said to be stale if the registration’s last update check time is non-null and the time difference in seconds calculated by the current time minus the registration’s last update check time is greater than 86400.

A service worker registration has an associated update via cache mode, which is "imports", "all", or "none". It is initially set to "imports".

A service worker registration has one or more task queues that back up the tasks from its active worker’s event loop’s corresponding task queues. (The target task sources for this back up operation are the handle fetch task source and the handle functional event task source.) The user agent dumps the active worker’s tasks to the service worker registration's task queues when the active worker is terminated and re-queues those tasks to the active worker’s event loop’s corresponding task queues when the active worker spins off. Unlike the task queues owned by event loops, the service worker registration's task queues are not processed by any event loops in and of itself.

A service worker registration has an associated NavigationPreloadManager object.

A service worker registration has an associated navigation preload enabled flag. It is initially unset.

A service worker registration has an associated navigation preload header value, which is a byte sequence. It is initially set to `true`.

A service worker registration is said to be unregistered if registration map[this service worker registration's (storage key, serialized scope url)] is not this service worker registration.

2.3.1. Lifetime

A user agent must persistently keep a list of registered service worker registrations unless otherwise they are explicitly unregistered. A user agent has a registration map that stores the entries of the tuple of service worker registration's (storage key, serialized scope url) and the corresponding service worker registration. The lifetime of service worker registrations is beyond that of the ServiceWorkerRegistration objects which represent them within the lifetime of their corresponding service worker clients.

2.4. Service Worker Client

A service worker client is an environment.

A service worker client has an associated discarded flag. It is initially unset.

Each service worker client has the following environment discarding steps:

Set client’s discarded flag.

Note: Implementations can discard clients whose discarded flag is set.

A service worker client has an algorithm defined as the origin that returns the service worker client's origin if the service worker client is an environment settings object, and the service worker client's creation URL’s origin otherwise.

A window client is a service worker client whose global object is a Window object.

A dedicated worker client is a service worker client whose global object is a DedicatedWorkerGlobalScope object.

A shared worker client is a service worker client whose global object is a SharedWorkerGlobalScope object.

A worker client is either a dedicated worker client or a shared worker client.

2.5. Control and Use

A service worker client has an active service worker that serves its own loading and its subresources. When a service worker client has a non-null active service worker, it is said to be controlled by that active service worker. When a service worker client is controlled by a service worker, it is said that the service worker client is using the service worker’s containing service worker registration. A service worker client's active service worker is determined as explained in the following subsections.

The rest of the section is non-normative.

The behavior in this section is not fully specified yet and will be specified in HTML Standard. The work is tracked by the issue and the pull request.

2.5.1. The window client case

A window client is created when a browsing context is created and when it navigates.

When a window client is created in the process of a browsing context creation:

If the browsing context's initial active document's origin is an opaque origin, the window client's active service worker is set to null. Otherwise, it is set to the creator document's service worker client's active service worker.

When a window client is created in the process of the browsing context's navigation:

If the fetch is routed through HTTP fetch, the window client's active service worker is set to the result of the service worker registration matching. Otherwise, if the created document's origin is an opaque origin or not the same as its creator document's origin, the window client's active service worker is set to null. Otherwise, it is set to the creator document's service worker client's active service worker.

Note: For an initial navigation with replacement enabled, the initial window client that was created when the browsing context was created is reused, but the active service worker is determined by the same behavior as above.

Note: Sandboxed iframes without the sandboxing directives, allow-same-origin and allow-scripts, result in having the active service worker value of null as their origin is an opaque origin.

2.5.2. The worker client case

A worker client is created when the user agent runs a worker.

When the worker client is created:

When the fetch is routed through HTTP fetch, the worker client's active service worker is set to the result of the service worker registration matching. Otherwise, if the worker client's origin is an opaque origin, or the request's URL is a blob URL and the worker client's origin is not the same as the origin of the last item in the worker client's global object's owner set, the worker client's active service worker is set to null. Otherwise, it is set to the active service worker of the environment settings object of the last item in the worker client's global object's owner set.

Note: Window clients and worker clients with a data: URL result in having the active service worker value of null as their origin is an opaque origin. Window clients and worker clients with a blob URL can inherit the active service worker of their creator document or owner, but if the request's origin is not the same as the origin of their creator document or owner, the active service worker is set to null.

2.6. Task Sources

The following additional task sources are used by service workers.

The handle fetch task source

This task source is used for dispatching fetch events to service workers.

The handle functional event task source

This task source is used for features that dispatch other functional events, e.g. push events, to service workers.

Note: A user agent may use a separate task source for each functional event type in order to avoid a head-of-line blocking phenomenon for certain functional events.

2.7. User Agent Shutdown

A user agent must maintain the state of its stored service worker registrations across restarts with the following rules:

An installing worker does not persist but is discarded. If the installing worker was the only service worker for the service worker registration, the service worker registration is discarded.
A waiting worker promotes to an active worker.

To attain this, the user agent must invoke Handle User Agent Shutdown when it terminates.

3. Client Context

Bootstrapping with a service worker:

// scope defaults to the path the script sits in
// "/" in this example
navigator.serviceWorker.register("/serviceworker.js").then(registration => {
  console.log("success!");
  if (registration.installing) {
    registration.installing.postMessage("Howdy from your installing page.");
  }
}, err => {
  console.error("Installing the worker failed!", err);
});

3.1. `ServiceWorker`

ServiceWorker

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

[SecureContext, Exposed=(Window,Worker)]
interface ServiceWorker : EventTarget {
  readonly attribute USVString scriptURL;
  readonly attribute ServiceWorkerState state;
  undefined postMessage(any message, sequence<object> transfer);
  undefined postMessage(any message, optional StructuredSerializeOptions options = {});

  // event
  attribute EventHandler onstatechange;
};
ServiceWorker includes AbstractWorker;

enum ServiceWorkerState {
  "parsed",
  "installing",
  "installed",
  "activating",
  "activated",
  "redundant"
};

A ServiceWorker object represents a service worker. Each ServiceWorker object is associated with a service worker. Multiple separate objects implementing the ServiceWorker interface across documents and workers can all be associated with the same service worker simultaneously.

A ServiceWorker object has an associated ServiceWorkerState object which is itself associated with service worker's state.

3.1.1. Getting `ServiceWorker` instances

An environment settings object has a service worker object map, a map where the keys are service workers and the values are ServiceWorker objects.

To get the service worker object representing serviceWorker (a service worker) in environment (an environment settings object), run these steps:

Let objectMap be environment’s service worker object map.
If objectMap[serviceWorker] does not exist, then:
1. Let serviceWorkerObj be a new ServiceWorker in environment’s Realm, and associate it with serviceWorker.
2. Set serviceWorkerObj’s state to serviceWorker’s state.
3. Set objectMap[serviceWorker] to serviceWorkerObj.
Return objectMap[serviceWorker].

3.1.2. `scriptURL`

ServiceWorker/scriptURL

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

The scriptURL getter steps are to return the service worker's serialized script url.

For example, consider a document created by a navigation to https://example.com/app.html which matches via the following registration call which has been previously executed:

// Script on the page https://example.com/app.html
navigator.serviceWorker.register("/service_worker.js");

The value of navigator.serviceWorker.controller.scriptURL will be "https://example.com/service_worker.js".

3.1.3. `state`

ServiceWorker/state

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

The state attribute must return the value (in ServiceWorkerState enumeration) to which it was last set.

3.1.4. `postMessage(message, transfer)`

The postMessage(message, transfer) method steps are:

Let options be «[ "transfer" → transfer ]».
Invoke postMessage(message, options) with message and options as the arguments.

3.1.5. `postMessage(message, options)`

The postMessage(message, options) method steps are:

Let serviceWorker be the service worker represented by this.
Let incumbentSettings be the incumbent settings object.
Let incumbentGlobal be incumbentSettings’s global object.
Let serializeWithTransferResult be StructuredSerializeWithTransfer(message, options["transfer"]). Rethrow any exceptions.
If the result of running the Should Skip Event algorithm with "message" and serviceWorker is true, then return.
Run these substeps in parallel:
1. If the result of running the Run Service Worker algorithm with serviceWorker is failure, then return.
2. Queue a task on the DOM manipulation task source to run the following steps:
  1. Let source be determined by switching on the type of incumbentGlobal:
    
    ServiceWorkerGlobalScope
    The result of getting the service worker object that represents incumbentGlobal’s service worker in the relevant settings object of serviceWorker’s global object.
    Window
    a new WindowClient object that represents incumbentGlobal’s relevant settings object.
    Otherwise
    a new Client object that represents incumbentGlobal’s associated worker
  2. Let origin be the serialization of incumbentSettings’s origin.
  3. Let destination be the ServiceWorkerGlobalScope object associated with serviceWorker.
  4. Let deserializeRecord be StructuredDeserializeWithTransfer(serializeWithTransferResult, destination’s Realm).
    
    If this throws an exception, let e be the result of creating an event named messageerror, using ExtendableMessageEvent, with the origin attribute initialized to origin and the source attribute initialized to source.
  5. Else:
    1. Let messageClone be deserializeRecord.[[Deserialized]].
    2. Let newPorts be a new frozen array consisting of all MessagePort objects in deserializeRecord.[[TransferredValues]], if any, maintaining their relative order.
    3. Let e be the result of creating an event named message, using ExtendableMessageEvent, with the origin attribute initialized to origin, the source attribute initialized to source, the data attribute initialized to messageClone, and the ports attribute initialized to newPorts.
  6. Dispatch e at destination.
  7. Invoke Update Service Worker Extended Events Set with serviceWorker and e.

3.1.6. Event handler

The following is the event handler (and its corresponding event handler event type) that must be supported, as event handler IDL attributes, by all objects implementing ServiceWorker interface:

event handler	event handler event type
ServiceWorker/statechange_event In all current engines. Firefox44+Safari11.1+Chrome40+ Opera27+Edge79+ Edge (Legacy)17+IENone Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+ `onstatechange`	`statechange`

3.2. `ServiceWorkerRegistration`

ServiceWorkerRegistration

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

[SecureContext, Exposed=(Window,Worker)]
interface ServiceWorkerRegistration : EventTarget {
  readonly attribute ServiceWorker? installing;
  readonly attribute ServiceWorker? waiting;
  readonly attribute ServiceWorker? active;
  [SameObject] readonly attribute NavigationPreloadManager navigationPreload;

  readonly attribute USVString scope;
  readonly attribute ServiceWorkerUpdateViaCache updateViaCache;

  [NewObject] Promise<undefined> update();
  [NewObject] Promise<boolean> unregister();

  // event
  attribute EventHandler onupdatefound;
};

enum ServiceWorkerUpdateViaCache {
  "imports",
  "all",
  "none"
};

A ServiceWorkerRegistration has a service worker registration (a service worker registration).

3.2.1. Getting `ServiceWorkerRegistration` instances

An environment settings object has a service worker registration object map, a map where the keys are service worker registrations and the values are ServiceWorkerRegistration objects.

To get the service worker registration object representing registration (a service worker registration) in environment (an environment settings object), run these steps:

Let objectMap be environment’s service worker registration object map.
If objectMap[registration] does not exist, then:
1. Let registrationObject be a new ServiceWorkerRegistration in environment’s Realm.
2. Set registrationObject’s service worker registration to registration.
3. Set registrationObject’s installing attribute to null.
4. Set registrationObject’s waiting attribute to null.
5. Set registrationObject’s active attribute to null.
6. If registration’s installing worker is not null, then set registrationObject’s installing attribute to the result of getting the service worker object that represents registration’s installing worker in environment.
7. If registration’s waiting worker is not null, then set registrationObject’s waiting attribute to the result of getting the service worker object that represents registration’s waiting worker in environment.
8. If registration’s active worker is not null, then set registrationObject’s active attribute to the result of getting the service worker object that represents registration’s active worker in environment.
9. Set objectMap[registration] to registrationObject.
Return objectMap[registration].

3.2.2. `installing`

ServiceWorkerRegistration/installing

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

installing attribute must return the value to which it was last set.

Note: Within a Realm, there is only one ServiceWorker object per associated service worker.

3.2.3. `waiting`

ServiceWorkerRegistration/waiting

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

waiting attribute must return the value to which it was last set.

Note: Within a Realm, there is only one ServiceWorker object per associated service worker.

3.2.4. `active`

ServiceWorkerRegistration/active

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

active attribute must return the value to which it was last set.

Note: Within a Realm, there is only one ServiceWorker object per associated service worker.

3.2.5. `navigationPreload`

ServiceWorkerRegistration/navigationPreload

In all current engines.

Firefox99+Safari15.4+Chrome59+

Opera46+Edge79+

Edge (Legacy)17+IENone

Firefox for Android99+iOS Safari15.4+Chrome for Android59+Android WebView59+Samsung Internet4.0+Opera Mobile43+

The navigationPreload getter steps are to return the service worker registration's NavigationPreloadManager object.

3.2.6. `scope`

ServiceWorkerRegistration/scope

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

The scope getter steps are to return the service worker registration's serialized scope url.

In the example in § 3.1.2 scriptURL, the value of registration.scope, obtained from navigator.serviceWorker.ready.then(registration => console.log(registration.scope)) for example, will be "https://example.com/".

3.2.7. `updateViaCache`

ServiceWorkerRegistration/updateViaCache

In all current engines.

Firefox57+Safari11.1+Chrome68+

Opera55+Edge79+

Edge (Legacy)18IENone

Firefox for Android57+iOS Safari11.3+Chrome for Android68+Android WebView68+Samsung Internet10.0+Opera Mobile48+

The updateViaCache getter steps are to return the service worker registration's update via cache mode.

3.2.8. `update()`

ServiceWorkerRegistration/update

In all current engines.

Firefox44+Safari11.1+Chrome45+

Opera32+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android45+Android WebView45+Samsung Internet4.0+Opera Mobile32+

The update() method steps are:

Let registration be the service worker registration.
Let newestWorker be the result of running Get Newest Worker algorithm passing registration as its argument.
If newestWorker is null, return a promise rejected with an "InvalidStateError" DOMException and abort these steps.
If this's relevant global object globalObject is a ServiceWorkerGlobalScope object, and globalObject’s associated service worker's state is "installing", return a promise rejected with an "InvalidStateError" DOMException and abort these steps.
Let promise be a promise.
Let job be the result of running Create Job with update, registration’s storage key, registration’s scope url, newestWorker’s script url, promise, and this's relevant settings object.
Set job’s worker type to newestWorker’s type.
Invoke Schedule Job with job.
Return promise.

3.2.9. `unregister()`

ServiceWorkerRegistration/unregister

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

Note: The unregister() method unregisters the service worker registration. It is important to note that the currently controlled service worker client's active service worker's containing service worker registration is effective until all the service worker clients (including itself) using this service worker registration unload. That is, the unregister() method only affects subsequent navigations.

The unregister() method steps are:

Let registration be the service worker registration.
Let promise be a new promise.
Let job be the result of running Create Job with unregister, registration’s storage key, registration’s scope url, null, promise, and this's relevant settings object.
Invoke Schedule Job with job.
Return promise.

3.2.10. Event handler

The following is the event handler (and its corresponding event handler event type) that must be supported, as event handler IDL attributes, by all objects implementing ServiceWorkerRegistration interface:

event handler	event handler event type
ServiceWorkerRegistration/updatefound_event In all current engines. Firefox44+Safari11.1+Chrome40+ Opera27+Edge79+ Edge (Legacy)17+IENone Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+ `onupdatefound`	`updatefound`

3.3. `navigator.serviceWorker`

partial interface Navigator {
  [SecureContext, SameObject] readonly attribute ServiceWorkerContainer serviceWorker;
};

partial interface WorkerNavigator {
  [SecureContext, SameObject] readonly attribute ServiceWorkerContainer serviceWorker;
};

Navigator/serviceWorker

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

The serviceWorker getter steps are to return the ServiceWorkerContainer object that is associated with this.

3.4. `ServiceWorkerContainer`

ServiceWorkerContainer

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

[SecureContext, Exposed=(Window,Worker)]
interface ServiceWorkerContainer : EventTarget {
  readonly attribute ServiceWorker? controller;
  readonly attribute Promise<ServiceWorkerRegistration> ready;

  [NewObject] Promise<ServiceWorkerRegistration> register(USVString scriptURL, optional RegistrationOptions options = {});

  [NewObject] Promise<(ServiceWorkerRegistration or undefined)> getRegistration(optional USVString clientURL = "");
  [NewObject] Promise<FrozenArray<ServiceWorkerRegistration>> getRegistrations();

  undefined startMessages();


  // events
  attribute EventHandler oncontrollerchange;
  attribute EventHandler onmessage; // event.source of message events is ServiceWorker object
  attribute EventHandler onmessageerror;
};

dictionary RegistrationOptions {
  USVString scope;
  WorkerType type = "classic";
  ServiceWorkerUpdateViaCache updateViaCache = "imports";
};

The user agent must create a ServiceWorkerContainer object when a Navigator object or a WorkerNavigator object is created and associate it with that object.

A ServiceWorkerContainer provides capabilities to register, unregister, and update the service worker registrations, and provides access to the state of the service worker registrations and their associated service workers.

A ServiceWorkerContainer has an associated service worker client, which is a service worker client whose global object is associated with the Navigator object or the WorkerNavigator object that the ServiceWorkerContainer is retrieved from.

A ServiceWorkerContainer object has an associated ready promise (a promise or null). It is initially null.

A ServiceWorkerContainer object has a task source called the client message queue, initially empty. A client message queue can be enabled or disabled, and is initially disabled. When a ServiceWorkerContainer object’s client message queue is enabled, the event loop must use it as one of its task sources. When the ServiceWorkerContainer object’s relevant global object is a Window object, all tasks queued on its client message queue must be associated with its relevant settings object’s responsible document.

3.4.1. `controller`

ServiceWorkerContainer/controller

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

controller attribute must run these steps:

Let client be this's service worker client.
If client’s active service worker is null, then return null.
Return the result of getting the service worker object that represents client’s active service worker in this's relevant settings object.

Note: navigator.serviceWorker.controller returns null if the request is a force refresh (shift+refresh).

3.4.2. `ready`

ServiceWorkerContainer/ready

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

ready attribute must run these steps:

If this's ready promise is null, then set this's ready promise to a new promise.
Let readyPromise be this's ready promise.
If readyPromise is pending, run the following substeps in parallel:
1. Let client by this's service worker client.
2. Let storage key be the result of running obtain a storage key given client.
3. Let registration be the result of running Match Service Worker Registration given storage key and client’s creation URL.
4. If registration is not null, and registration’s active worker is not null, queue a task on readyPromise’s relevant settings object's responsible event loop, using the DOM manipulation task source, to resolve readyPromise with the result of getting the service worker registration object that represents registration in readyPromise’s relevant settings object.
Return readyPromise.

Note: The returned ready promise will never reject. If it does not resolve in this algorithm, it will eventually resolve when a matching service worker registration is registered and its active worker is set. (See the relevant Activate algorithm step.)

3.4.3. `register(scriptURL, options)`

ServiceWorkerContainer/register

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

Note: The register(scriptURL, options) method creates or updates a service worker registration for the given scope url. If successful, a service worker registration ties the provided scriptURL to a scope url, which is subsequently used for navigation matching.

The register(scriptURL, options) method steps are:

Let p be a promise.
Let client be this's service worker client.
Let scriptURL be the result of parsing scriptURL with this's relevant settings object’s API base URL.
Let scopeURL be null.
If options["scope"] exists, set scopeURL to the result of parsing options["scope"] with this's relevant settings object’s API base URL.
Invoke Start Register with scopeURL, scriptURL, p, client, client’s creation URL, options["type"], and options["updateViaCache"].
Return p.

3.4.4. `getRegistration(clientURL)`

ServiceWorkerContainer/getRegistration

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

getRegistration(clientURL) method steps are:

Let client be this's service worker client.
Let storage key be the result of running obtain a storage key given client.
Let clientURL be the result of parsing clientURL with this's relevant settings object’s API base URL.
If clientURL is failure, return a promise rejected with a TypeError.
Set clientURL’s fragment to null.
If the origin of clientURL is not client’s origin, return a promise rejected with a "SecurityError" DOMException.
Let promise be a new promise.
Run the following substeps in parallel:
1. Let registration be the result of running Match Service Worker Registration given storage key and clientURL.
2. If registration is null, resolve promise with undefined and abort these steps.
3. Resolve promise with the result of getting the service worker registration object that represents registration in promise’s relevant settings object.
Return promise.

3.4.5. `getRegistrations()`

ServiceWorkerContainer/getRegistrations

In all current engines.

Firefox44+Safari11.1+Chrome45+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android45+Android WebView40+Samsung Internet4.0+Opera Mobile27+

getRegistrations() method steps are:

Let client be this's service worker client.
Let client storage key be the result of running obtain a storage key given client.
Let promise be a new promise.
Run the following steps in parallel:
1. Let registrations be a new list.
2. For each (storage key, scope) → registration of registration map:
  1. If storage key equals client storage key, then append registration to registrations.
3. Queue a task on promise’s relevant settings object's responsible event loop, using the DOM manipulation task source, to run the following steps:
  1. Let registrationObjects be a new list.
  2. For each registration of registrations:
    1. Let registrationObj be the result of getting the service worker registration object that represents registration in promise’s relevant settings object.
    2. Append registrationObj to registrationObjects.
  3. Resolve promise with a new frozen array of registrationObjects in promise’s relevant Realm.
Return promise.

3.4.6. `startMessages()`

ServiceWorkerContainer/startMessages

In all current engines.

Firefox64+Safari11.1+Chrome74+

Opera62+Edge79+

Edge (Legacy)NoneIENone

Firefox for Android64+iOS Safari11.3+Chrome for Android74+Android WebView74+Samsung Internet11.0+Opera Mobile50+

The startMessages() method steps are to enable this's client message queue if it is not enabled.

3.4.7. Event handlers

The following are the event handlers (and their corresponding event handler event types) that must be supported, as event handler IDL attributes, by all objects implementing the ServiceWorkerContainer interface:

event handler	event handler event type
ServiceWorkerContainer/controllerchange_event In all current engines. Firefox44+Safari11.1+Chrome40+ Opera27+Edge79+ Edge (Legacy)17+IENone Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+ `oncontrollerchange`	`controllerchange`
ServiceWorkerContainer/message_event In all current engines. Firefox44+Safari11.1+Chrome40+ Opera27+Edge79+ Edge (Legacy)17+IENone Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+ `onmessage`	`message`
`onmessageerror`	`messageerror`

The first time the onmessage setter steps are performed, enable this's client message queue.

3.5. Events

The following event is dispatched on ServiceWorker object:

Event name	Interface	Dispatched when…
`statechange`	`Event`	The `state` attribute of the `ServiceWorker` object is changed.

The following event is dispatched on ServiceWorkerRegistration object:

Event name	Interface	Dispatched when…
`updatefound`	`Event`	The service worker registration's installing worker changes. (See step 8 of the Install algorithm.)

The following events are dispatched on ServiceWorkerContainer object:

Event name	Interface	Dispatched when…
`controllerchange`	`Event`	The service worker client's active service worker changes. (See step 9.2 of the Activate algorithm. The skip waiting flag of a service worker causes activation of the service worker registration to occur while service worker clients are using the service worker registration, `navigator.serviceWorker.controller` immediately reflects the active worker as the service worker that controls the service worker client.)
`message`	`Event`	The service worker client receives a message from a service worker. See `postMessage(message, options)`.
`messageerror`	`Event`	The service worker client is sent a message that cannot be deserialized from a service worker. See `postMessage(message, options)`.

NavigationPreloadManager

In all current engines.

Firefox99+Safari15.4+Chrome59+

Opera46+Edge79+

Edge (Legacy)18IENone

Firefox for Android99+iOS Safari15.4+Chrome for Android59+Android WebView59+Samsung Internet7.0+Opera Mobile43+

[SecureContext, Exposed=(Window,Worker)]
interface NavigationPreloadManager {
  Promise<undefined> enable();
  Promise<undefined> disable();
  Promise<undefined> setHeaderValue(ByteString value);
  Promise<NavigationPreloadState> getState();
};

dictionary NavigationPreloadState {
  boolean enabled = false;
  ByteString headerValue;
};

NavigationPreloadManager/enable

In all current engines.

Firefox99+Safari15.4+Chrome59+

Opera46+Edge79+

Edge (Legacy)18IENone

Firefox for Android99+iOS Safari15.4+Chrome for Android59+Android WebView59+Samsung Internet7.0+Opera Mobile43+

The enable() method steps are:

Let promise be a new promise.
Run the following steps in parallel:
Let registration be this's associated service worker registration.
If registration’s active worker is null, reject promise with an "InvalidStateError" DOMException, and abort these steps.
Set registration’s navigation preload enabled flag.
Resolve promise with undefined.
Return promise.

NavigationPreloadManager/disable

In all current engines.

Firefox99+Safari15.4+Chrome59+

Opera46+Edge79+

Edge (Legacy)18IENone

Firefox for Android99+iOS Safari15.4+Chrome for Android59+Android WebView59+Samsung Internet7.0+Opera Mobile43+

The disable() method steps are:

Let promise be a new promise.
Run the following steps in parallel:
Let registration be this's associated service worker registration.
If registration’s active worker is null, reject promise with an "InvalidStateError" DOMException, and abort these steps.
Unset registration’s navigation preload enabled flag.
Resolve promise with undefined.
Return promise.

NavigationPreloadManager/setHeaderValue

In all current engines.

Firefox99+Safari15.4+Chrome59+

Opera46+Edge79+

Edge (Legacy)18IENone

Firefox for Android99+iOS Safari15.4+Chrome for Android59+Android WebView59+Samsung Internet7.0+Opera Mobile43+

The setHeaderValue(value) method steps are:

Let promise be a new promise.
Run the following steps in parallel:
Let registration be this's associated service worker registration.
If registration’s active worker is null, reject promise with an "InvalidStateError" DOMException, and abort these steps.
Set registration’s navigation preload header value to value.
Resolve promise with undefined.
Return promise.

NavigationPreloadManager/getState

In all current engines.

Firefox99+Safari15.4+Chrome59+

Opera46+Edge79+

Edge (Legacy)18IENone

Firefox for Android99+iOS Safari15.4+Chrome for Android59+Android WebView59+Samsung Internet7.0+Opera Mobile43+

The getState() method steps are:

Let promise be a new promise.
Run the following steps in parallel:
Let registration be this's associated service worker registration.
Let state be a new NavigationPreloadState dictionary.
If registration’s navigation preload enabled flag is set, set state["enabled"] to true.
Set state["headerValue"] to registration’s navigation preload header value.
Resolve promise with state.
Return promise.

4. Execution Context

Serving Cached Resources:

// caching.js
self.addEventListener("install", event => {
  event.waitUntil(
    // Open a cache of resources.
    caches.open("shell-v1").then(cache => {
      // Begins the process of fetching them. Succeeds only once all
      // resources have been stored. Even just one failing resource
      // causes the entire operation to fail.
      return cache.addAll([
        "/app.html",
        "/assets/v1/base.css",
        "/assets/v1/app.js",
        "/assets/v1/logo.png",
        "/assets/v1/intro_video.webm"
      ]);
    })
  );
});

self.addEventListener("fetch", event => {
  // No "fetch" events are dispatched to the service worker until it
  // successfully installs and activates.

  // All operations on caches are async, including matching URLs, so we use
  // promises heavily. e.respondWith() even takes promises to enable this:
  event.respondWith(
    caches.match(e.request).then(response => {
      return response || fetch(e.request);
    }).catch(() => {
      return caches.match("/fallback.html");
    })
  );
});

4.1. `ServiceWorkerGlobalScope`

ServiceWorkerGlobalScope

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera24+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile24+

[Global=(Worker,ServiceWorker), Exposed=ServiceWorker]
interface ServiceWorkerGlobalScope : WorkerGlobalScope {
  [SameObject] readonly attribute Clients clients;
  [SameObject] readonly attribute ServiceWorkerRegistration registration;
  [SameObject] readonly attribute ServiceWorker serviceWorker;

  [NewObject] Promise<undefined> skipWaiting();

  attribute EventHandler oninstall;
  attribute EventHandler onactivate;
  attribute EventHandler onfetch;

  attribute EventHandler onmessage;
  attribute EventHandler onmessageerror;
};

A ServiceWorkerGlobalScope object represents the global execution context of a service worker. A ServiceWorkerGlobalScope object has an associated service worker (a service worker). A ServiceWorkerGlobalScope object has an associated force bypass cache for import scripts flag. It is initially unset.

Note: ServiceWorkerGlobalScope object provides generic, event-driven, time-limited script execution contexts that run at an origin. Once successfully registered, a service worker is started, kept alive and killed by their relationship to events, not service worker clients. Any type of synchronous requests must not be initiated inside of a service worker.

4.1.1. `clients`

ServiceWorkerGlobalScope/clients

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera24+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile24+

The clients getter steps are to return the Clients object that is associated with this.

4.1.2. `registration`

ServiceWorkerGlobalScope/registration

In all current engines.

Firefox44+Safari11.1+Chrome42+

Opera26+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android42+Android WebView42+Samsung Internet4.0+Opera Mobile26+

The registration getter steps are to return the result of getting the service worker registration object representing this's service worker's containing service worker registration in this's relevant settings object.

4.1.3. `serviceWorker`

The serviceWorker getter steps are to return the result of getting the service worker object that represents this's service worker in this's relevant settings object.

4.1.4. `skipWaiting()`

ServiceWorkerGlobalScope/skipWaiting

In all current engines.

Firefox44+Safari11.1+Chrome41+

Opera25+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android41+Android WebView41+Samsung Internet4.0+Opera Mobile25+

Note: The skipWaiting() method allows this service worker to progress from the registration's waiting position to active even while service worker clients are using the registration.

The skipWaiting() method steps are:

Let promise be a new promise.
Run the following substeps in parallel:
1. Set service worker's skip waiting flag.
2. Invoke Try Activate with service worker's containing service worker registration.
3. Resolve promise with undefined.
Return promise.

4.1.5. Event handlers

event handler	event handler event type
ServiceWorkerGlobalScope/install_event In all current engines. Firefox44+Safari11.1+Chrome40+ Opera24+Edge79+ Edge (Legacy)17+IENone Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile24+ `oninstall`	`install`
ServiceWorkerGlobalScope/activate_event In all current engines. Firefox44+Safari11.1+Chrome40+ Opera24+Edge79+ Edge (Legacy)17+IENone Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile24+ `onactivate`	`activate`
ServiceWorkerGlobalScope/fetch_event In all current engines. Firefox44+Safari11.1+Chrome40+ Opera24+Edge79+ Edge (Legacy)17+IENone Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile24+ `onfetch`	`fetch`
ServiceWorkerGlobalScope/message_event In all current engines. Firefox44+Safari11.1+Chrome40+ Opera24+Edge79+ Edge (Legacy)17+IENone Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile24+ `onmessage`	`message`
`onmessageerror`	`messageerror`

4.2. `Client`

Client

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

WindowClient

In all current engines.

Firefox44+Safari11.1+Chrome42+

Opera29+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android42+Android WebView42+Samsung Internet4.0+Opera Mobile29+

[Exposed=ServiceWorker]
interface Client {
  readonly attribute USVString url;
  readonly attribute FrameType frameType;
  readonly attribute DOMString id;
  readonly attribute ClientType type;
  undefined postMessage(any message, sequence<object> transfer);
  undefined postMessage(any message, optional StructuredSerializeOptions options = {});
};

[Exposed=ServiceWorker]
interface WindowClient : Client {
  readonly attribute VisibilityState visibilityState;
  readonly attribute boolean focused;
  [SameObject] readonly attribute FrozenArray<USVString> ancestorOrigins;
  [NewObject] Promise<WindowClient> focus();
  [NewObject] Promise<WindowClient?> navigate(USVString url);
};

enum FrameType {
  "auxiliary",
  "top-level",
  "nested",
  "none"
};

A Client object has an associated service worker client (a service worker client).

A Client object has an associated frame type, which is one of "auxiliary", "top-level", "nested", and "none". Unless stated otherwise it is "none".

A WindowClient object has an associated browsing context, which is its service worker client's global object's browsing context.

A WindowClient object has an associated visibility state, which is one of visibilityState attribute value.

A WindowClient object has an associated focus state, which is either true or false (initially false).

A WindowClient object has an associated ancestor origins array.

4.2.1. `url`

Client/url

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

The url getter steps are to return this's associated service worker client's serialized creation URL.

4.2.2. `frameType`

Client/frameType

In all current engines.

Firefox44+Safari11.1+Chrome43+

Opera30+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android43+Android WebView43+Samsung Internet4.0+Opera Mobile30+

The frameType getter steps are to return this's frame type.

4.2.3. `id`

Client/id

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

The id getter steps are to return this's associated service worker client's id.

4.2.4. `type`

Client/type

In all current engines.

Firefox54+Safari11.1+Chrome60+

Opera47+Edge79+

Edge (Legacy)17+IENone

Firefox for Android54+iOS Safari11.3+Chrome for Android60+Android WebView60+Samsung Internet8.0+Opera Mobile44+

The type getter steps are:

Let client be this's service worker client.
If client is an environment settings object, then:
1. If client is a window client, return "window".
2. Else if client is a dedicated worker client, return "worker".
3. Else if client is a shared worker client, return "sharedworker".
Else:
1. Return "window".

4.2.5. `postMessage(message, transfer)`

The postMessage(message, transfer) method steps are:

Let options be «[ "transfer" → transfer ]».
Invoke postMessage(message, options) with message and options as the arguments.

4.2.6. `postMessage(message, options)`

Client/postMessage

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

The postMessage(message, options) method steps are:

Let contextObject be this.
Let sourceSettings be the contextObject’s relevant settings object.
Let serializeWithTransferResult be StructuredSerializeWithTransfer(message, options["transfer"]). Rethrow any exceptions.
Run the following steps in parallel:
1. Let targetClient be null.
2. For each service worker client client:
  1. If client is the contextObject’s service worker client, set targetClient to client, and break.
3. If targetClient is null, return.
4. Let destination be the ServiceWorkerContainer object whose associated service worker client is targetClient.
5. Add a task that runs the following steps to destination’s client message queue:
  1. Let origin be the serialization of sourceSettings’s origin.
  2. Let source be the result of getting the service worker object that represents contextObject’s relevant global object's service worker in targetClient.
  3. Let deserializeRecord be StructuredDeserializeWithTransfer(serializeWithTransferResult, destination’s relevant Realm).
    
    If this throws an exception, catch it, fire an event named messageerror at destination, using MessageEvent, with the origin attribute initialized to origin and the source attribute initialized to source, and then abort these steps.
  4. Let messageClone be deserializeRecord.[[Deserialized]].
  5. Let newPorts be a new frozen array consisting of all MessagePort objects in deserializeRecord.[[TransferredValues]], if any.
  6. Dispatch an event named message at destination, using MessageEvent, with the origin attribute initialized to origin, the source attribute initialized to source, the data attribute initialized to messageClone, and the ports attribute initialized to newPorts.

4.2.7. `visibilityState`

WindowClient/visibilityState

In all current engines.

Firefox44+Safari11.1+Chrome42+

Opera29+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android42+Android WebView42+Samsung Internet4.0+Opera Mobile29+

The visibilityState getter steps are to return this's visibility state.

4.2.8. `focused`

WindowClient/focused

In all current engines.

Firefox44+Safari11.1+Chrome42+

Opera29+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android42+Android WebView42+Samsung Internet4.0+Opera Mobile29+

The focused getter steps are to return this's focus state.

4.2.9. `ancestorOrigins`

The ancestorOrigins getter steps are to return this's associated ancestor origins array.

4.2.10. `focus()`

WindowClient/focus

In all current engines.

Firefox44+Safari11.1+Chrome42+

Opera29+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android42+Android WebView42+Samsung Internet4.0+Opera Mobile29+

The focus() method steps are:

If this algorithm is not triggered by user activation, return a promise rejected with an "InvalidAccessError" DOMException.
Let serviceWorkerEventLoop be the surrounding agent's event loop.
Let promise be a new promise.
Queue a task to run the following steps on this's associated service worker client's responsible event loop using the user interaction task source:
1. Run the focusing steps with this's browsing context.
2. Let frameType be the result of running Get Frame Type with this's browsing context.
3. Let visibilityState be this's browsing context's active document's visibilityState attribute value.
4. Let focusState be the result of running the has focus steps with this's browsing context's active document.
5. Let ancestorOriginsList be this's browsing context's active document's relevant global object's Location object’s ancestor origins list's associated list.
6. Queue a task to run the following steps on serviceWorkerEventLoop using the DOM manipulation task source:
  1. Let windowClient be the result of running Create Window Client with this's associated service worker client, frameType, visibilityState, focusState, and ancestorOriginsList.
  2. If windowClient’s focus state is true, resolve promise with windowClient.
  3. Else, reject promise with a TypeError.
Return promise.

4.2.11. `navigate(url)`

WindowClient/navigate

Firefox44+SafariNoneChrome49+

Opera36+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS SafariNoneChrome for Android49+Android WebView42+Samsung Internet4.0+Opera Mobile36+

The navigate(url) method steps are:

Let url be the result of parsing url with this's relevant settings object’s API base URL.
If url is failure, return a promise rejected with a TypeError.
If url is about:blank, return a promise rejected with a TypeError.
If this's associated service worker client's active service worker is not this's relevant global object’s service worker, return a promise rejected with a TypeError.
Let serviceWorkerEventLoop be the current global object's event loop.
Let promise be a new promise.
Queue a task to run the following steps on this's associated service worker client's responsible event loop using the user interaction task source:
1. Let browsingContext be this's browsing context.
2. If browsingContext has discarded its Document, queue a task to reject promise with a TypeError, on serviceWorkerEventLoop using the DOM manipulation task source, and abort these steps.
3. HandleNavigate: Navigate browsingContext to url with exceptions enabled. The source browsing context must be browsingContext.
4. If the algorithm steps invoked in the step labeled HandleNavigate throws an exception, queue a task to reject promise with the exception, on serviceWorkerEventLoop using the DOM manipulation task source, and abort these steps.
5. Let frameType be the result of running Get Frame Type with browsingContext.
6. Let visibilityState be browsingContext’s active document’s visibilityState attribute value.
7. Let focusState be the result of running the has focus steps with browsingContext’s active document.
8. Let ancestorOriginsList be browsingContext’s active document's relevant global object's Location object’s ancestor origins list's associated list.
9. Queue a task to run the following steps on serviceWorkerEventLoop using the DOM manipulation task source:
  1. If browsingContext’s Window object’s environment settings object’s creation URL’s origin is not the same as the service worker's origin, resolve promise with null and abort these steps.
  2. Let windowClient be the result of running Create Window Client with this's service worker client, frameType, visibilityState, focusState, and ancestorOriginsList.
  3. Resolve promise with windowClient.
Return promise.

4.3. `Clients`

Clients

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

[Exposed=ServiceWorker]
interface Clients {
  // The objects returned will be new instances every time
  [NewObject] Promise<(Client or undefined)> get(DOMString id);
  [NewObject] Promise<FrozenArray<Client>> matchAll(optional ClientQueryOptions options = {});
  [NewObject] Promise<WindowClient?> openWindow(USVString url);
  [NewObject] Promise<undefined> claim();
};

dictionary ClientQueryOptions {
  boolean includeUncontrolled = false;
  ClientType type = "window";
};

enum ClientType {
  "window",
  "worker",
  "sharedworker",
  "all"
};

The user agent must create a Clients object when a ServiceWorkerGlobalScope object is created and associate it with that object.

4.3.1. `get(id)`

Clients/get

In all current engines.

Firefox45+Safari11.1+Chrome51+

Opera38+Edge79+

Edge (Legacy)17+IENone

Firefox for Android45+iOS Safari11.3+Chrome for Android51+Android WebView51+Samsung Internet5.0+Opera Mobile41+

The get(id) method steps are:

Let promise be a new promise.
Run these substeps in parallel:
1. For each service worker client client where the result of running obtain a storage key given client equals the associated service worker's containing service worker registration's storage key:
  1. If client’s id is not id, continue.
  2. Wait for either client’s execution ready flag to be set or for client’s discarded flag to be set.
  3. If client’s execution ready flag is set, then invoke Resolve Get Client Promise with client and promise, and abort these steps.
2. Resolve promise with undefined.
Return promise.

4.3.2. `matchAll(options)`

Clients/matchAll

In all current engines.

Firefox54+Safari11.1+Chrome42+

Opera29+Edge79+

Edge (Legacy)17+IENone

Firefox for Android54+iOS Safari11.3+Chrome for Android42+Android WebView42+Samsung Internet4.0+Opera Mobile29+

The matchAll(options) method steps are:

Let promise be a new promise.
Run the following steps in parallel:
1. Let targetClients be a new list.
2. For each service worker client client where the result of running obtain a storage key given client equals the associated service worker's containing service worker registration's storage key:
  1. If client’s execution ready flag is unset or client’s discarded flag is set, continue.
  2. If client is not a secure context, continue.
  3. If options["includeUncontrolled"] is false, and if client’s active service worker is not the associated service worker, continue.
  4. Add client to targetClients.
3. Let matchedWindowData be a new list.
4. Let matchedClients be a new list.
5. For each service worker client client in targetClients:
  1. If options["type"] is "window" or "all", and client is not an environment settings object or is a window client, then:
    1. Let windowData be «[ "client" → client, "ancestorOriginsList" → a new list ]».
    2. Let browsingContext be null.
    3. Let isClientEnumerable be true.
    4. If client is an environment settings object, set browsingContext to client’s global object's browsing context.
    5. Else, set browsingContext to client’s target browsing context.
    6. Queue a task task to run the following substeps on browsingContext’s event loop using the user interaction task source:
      1. If browsingContext has been discarded, then set isClientEnumerable to false and abort these steps.
      2. If client is a window client and client’s responsible document is not browsingContext’s active document, then set isClientEnumerable to false and abort these steps.
      3. Set windowData["frameType"] to the result of running Get Frame Type with browsingContext.
      4. Set windowData["visibilityState"] to browsingContext’s active document's visibilityState attribute value.
      5. Set windowData["focusState"] to the result of running the has focus steps with browsingContext’s active document as the argument.
      6. If client is a window client, then set windowData["ancestorOriginsList"] to browsingContext’s active document's relevant global object's Location object’s ancestor origins list's associated list.
    7. Wait for task to have executed.
      
      Note: Wait is a blocking wait, but implementers may run the iterations in parallel as long as the state is not broken.
    8. If isClientEnumerable is true, then:
      1. Add windowData to matchedWindowData.
  2. Else if options["type"] is "worker" or "all" and client is a dedicated worker client, or options["type"] is "sharedworker" or "all" and client is a shared worker client, then:
    1. Add client to matchedClients.
6. Queue a task to run the following steps on promise’s relevant settings object's responsible event loop using the DOM manipulation task source:
  1. Let clientObjects be a new list.
  2. For each windowData in matchedWindowData:
    1. Let windowClient be the result of running Create Window Client algorithm with windowData["client"], windowData["frameType"], windowData["visibilityState"], windowData["focusState"], and windowData["ancestorOriginsList"] as the arguments.
    2. Append windowClient to clientObjects.
  3. For each client in matchedClients:
    1. Let clientObject be the result of running Create Client algorithm with client as the argument.
    2. Append clientObject to clientObjects.
  4. Sort clientObjects such that:
    - WindowClient objects whose browsing context has been focused are placed first, sorted in the most recently focused order.
    - WindowClient objects whose browsing context has never been focused are placed next, sorted in their service worker client's creation order.
    - Client objects whose associated service worker client is a worker client are placed next, sorted in their service worker client's creation order.
    Note: Window clients are always placed before worker clients.
  5. Resolve promise with a new frozen array of clientObjects in promise’s relevant Realm.
Return promise.

4.3.3. `openWindow(url)`

Clients/openWindow

In all current engines.

Firefox45+Safari11.1+Chrome51+

Opera38+Edge79+

Edge (Legacy)NoneIENone

Firefox for Android45+iOS Safari11.3+Chrome for Android51+Android WebView51+Samsung Internet5.0+Opera Mobile41+

The openWindow(url) method steps are:

Let url be the result of parsing url with this's relevant settings object’s API base URL.
If url is failure, return a promise rejected with a TypeError.
If url is about:blank, return a promise rejected with a TypeError.
If this algorithm is not triggered by user activation, return a promise rejected with an "InvalidAccessError" DOMException.
Let serviceWorkerEventLoop be the current global object's event loop.
Let promise be a new promise.
Run these substeps in parallel:
1. Let newContext be a new top-level browsing context.
2. Queue a task to run the following steps on newContext’s Window object’s environment settings object's responsible event loop using the user interaction task source:
  1. HandleNavigate: Navigate newContext to url with exceptions enabled and replacement enabled.
  2. If the algorithm steps invoked in the step labeled HandleNavigate throws an exception, queue a task to reject promise with the exception, on serviceWorkerEventLoop using the DOM manipulation task source, and abort these steps.
  3. Let frameType be the result of running Get Frame Type with newContext.
  4. Let visibilityState be newContext’s active document’s visibilityState attribute value.
  5. Let focusState be the result of running the has focus steps with newContext’s active document as the argument.
  6. Let ancestorOriginsList be newContext’s active document’s relevant global object’s Location object’s ancestor origins list's associated list.
  7. Queue a task to run the following steps on serviceWorkerEventLoop using the DOM manipulation task source:
    1. If the result of running obtain a storage key given newContext’s Window object’s environment settings object is not equal to the service worker's containing service worker registration's storage key, then resolve promise with null and abort these steps.
    2. Let client be the result of running Create Window Client with newContext’s Window object’s environment settings object, frameType, visibilityState, focusState, and ancestorOriginsList as the arguments.
    3. Resolve promise with client.
Return promise.

4.3.4. `claim()`

Clients/claim

In all current engines.

Firefox44+Safari11.1+Chrome42+

Opera29+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android42+Android WebView42+Samsung Internet4.0+Opera Mobile29+

The claim() method steps are:

If the service worker is not an active worker, return a promise rejected with an "InvalidStateError" DOMException.
Let promise be a new promise.
Run the following substeps in parallel:
1. For each service worker client client where the result of running obtain a storage key given client equals the service worker's containing service worker registration's storage key:
  1. If client’s execution ready flag is unset or client’s discarded flag is set, continue.
  2. If client is not a secure context, continue.
  3. Let storage key be the result of running obtain a storage key given client.
  4. Let registration be the result of running Match Service Worker Registration given storage key and client’s creation URL.
  5. If registration is not the service worker's containing service worker registration, continue.
    
    Note: registration will be null if the service worker's containing service worker registration is unregistered.
  6. If client’s active service worker is not the service worker, then:
    1. Invoke Handle Service Worker Client Unload with client as the argument.
    2. Set client’s active service worker to service worker.
    3. Invoke Notify Controller Change algorithm with client as the argument.
2. Resolve promise with undefined.
Return promise.

4.4. `ExtendableEvent`

ExtendableEvent

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera24+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile24+

ExtendableEvent/ExtendableEvent

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera24+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile24+

[Exposed=ServiceWorker]
interface ExtendableEvent : Event {
  constructor(DOMString type, optional ExtendableEventInit eventInitDict = {});
  undefined waitUntil(Promise<any> f);
};

dictionary ExtendableEventInit : EventInit {
  // Defined for the forward compatibility across the derived events
};

An ExtendableEvent object has an associated extend lifetime promises (an array of promises). It is initially an empty array.

An ExtendableEvent object has an associated pending promises count (the number of pending promises in the extend lifetime promises). It is initially set to zero.

An ExtendableEvent object has an associated timed out flag. It is initially unset, and is set after an optional user agent imposed delay if the pending promises count is greater than zero.

An ExtendableEvent object is said to be active when its timed out flag is unset and either its pending promises count is greater than zero or its dispatch flag is set.

Service workers have two lifecycle events, install and activate. Service workers use the ExtendableEvent interface for activate event and install event.

Service worker extensions that define event handlers may also use or extend the ExtendableEvent interface.

4.4.1. `event.waitUntil(f)`

Note: waitUntil() method extends the lifetime of the event.

ExtendableEvent/waitUntil

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera24+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile24+

The waitUntil(f) method steps are to add lifetime promise f to this.

To add lifetime promise promise (a promise) to event (an ExtendableEvent), run these steps:

If event’s isTrusted attribute is false, throw an "InvalidStateError" DOMException.
If event is not active, throw an "InvalidStateError" DOMException.

Note: If no lifetime extension promise has been added in the task that called the event handlers, calling waitUntil() in subsequent asynchronous tasks will throw.
Add promise to event’s extend lifetime promises.
Increment event’s pending promises count by one.

Note: The pending promises count is incremented even if the given promise has already been settled. The corresponding count decrement is done in the microtask queued by the reaction to the promise.
Upon fulfillment or rejection of promise, queue a microtask to run these substeps:
1. Decrement event’s pending promises count by one.
2. If event’s pending promises count is 0, then:
  1. Let registration be the current global object's associated service worker's containing service worker registration.
  2. If registration is unregistered, invoke Try Clear Registration with registration.
  3. If registration is not null, invoke Try Activate with registration.

The user agent should not terminate a service worker if Service Worker Has No Pending Events returns false for that service worker.

Service workers and extensions that define event handlers may define their own behaviors, allowing the extend lifetime promises to suggest operation length, and the rejected state of any of the promise in extend lifetime promises to suggest operation failure.

Note: Service workers delay treating the installing worker as "installed" until all the promises in the install event’s extend lifetime promises resolve successfully. (See the relevant Install algorithm step.) If any of the promises rejects, the installation fails. This is primarily used to ensure that a service worker is not considered "installed" until all of the core caches it depends on are populated. Likewise, service workers delay treating the active worker as "activated" until all the promises in the activate event’s extend lifetime promises settle. (See the relevant Activate algorithm step.) This is primarily used to ensure that any functional events are not dispatched to the service worker until it upgrades database schemas and deletes the outdated cache entries.

4.5. `FetchEvent`

FetchEvent

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

FetchEvent/FetchEvent

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

[Exposed=ServiceWorker]
interface FetchEvent : ExtendableEvent {
  constructor(DOMString type, FetchEventInit eventInitDict);
  [SameObject] readonly attribute Request request;
  readonly attribute Promise<any> preloadResponse;
  readonly attribute DOMString clientId;
  readonly attribute DOMString resultingClientId;
  readonly attribute DOMString replacesClientId;
  readonly attribute Promise<undefined> handled;

  undefined respondWith(Promise<Response> r);
};

dictionary FetchEventInit : ExtendableEventInit {
  required Request request;
  Promise<any> preloadResponse;
  DOMString clientId = "";
  DOMString resultingClientId = "";
  DOMString replacesClientId = "";
  Promise<undefined> handled;
};

Service workers have an essential functional event fetch. For fetch event, service workers use the FetchEvent interface which extends the ExtendableEvent interface.

Each event using FetchEvent interface has an associated potential response (a response), initially set to null, and the following associated flags that are initially unset:

wait to respond flag
respond-with entered flag
respond-with error flag

4.5.1. `event.request`

FetchEvent/request

In all current engines.

Firefox44+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

request attribute must return the value it was initialized to.

4.5.2. `event.preloadResponse`

FetchEvent/preloadResponse

FirefoxNoneSafari15.4+Chrome59+

Opera46+Edge79+

Edge (Legacy)18IENone

Firefox for AndroidNoneiOS Safari15.4+Chrome for Android59+Android WebView59+Samsung Internet7.0+Opera Mobile43+

preloadResponse attribute must return the value it was initialized to. When an event is created the attribute must be initialized to a promise resolved with undefined.

4.5.3. `event.clientId`

FetchEvent/clientId

In all current engines.

Firefox45+Safari11.1+Chrome49+

Opera36+Edge79+

Edge (Legacy)17+IENone

Firefox for Android45+iOS Safari11.3+Chrome for Android49+Android WebView49+Samsung Internet5.0+Opera Mobile36+

clientId attribute must return the value it was initialized to. When an event is created the attribute must be initialized to the empty string.

4.5.4. `event.resultingClientId`

FetchEvent/resultingClientId

In all current engines.

Firefox65+Safari16+Chrome72+

Opera60+Edge79+

Edge (Legacy)NoneIENone

Firefox for Android65+iOS Safari16+Chrome for Android72+Android WebView72+Samsung Internet11.0+Opera Mobile50+

resultingClientId attribute must return the value it was initialized to. When an event is created the attribute must be initialized to the empty string.

4.5.5. `event.replacesClientId`

FetchEvent/replacesClientId

In no current engines.

FirefoxNoneSafariNoneChromeNone

OperaNoneEdgeNone

Edge (Legacy)NoneIENone

Firefox for AndroidNoneiOS SafariNoneChrome for AndroidNoneAndroid WebViewNoneSamsung InternetNoneOpera MobileNone

replacesClientId attribute must return the value it was initialized to. When an event is created the attribute must be initialized to the empty string.

4.5.6. `event.handled`

handled attribute must return the value it was initialized to. When an event is created the attribute must be initialized to a pending promise.

4.5.7. `event.respondWith(r)`

FetchEvent/respondWith

In all current engines.

Firefox59+Safari11.1+Chrome42+

Opera29+Edge79+

Edge (Legacy)17+IENone

Firefox for Android59+iOS Safari11.3+Chrome for Android42+Android WebView42+Samsung Internet4.0+Opera Mobile29+

Note: Developers can set the argument r with either a promise that resolves with a Response object or a Response object (which is automatically cast to a promise). Otherwise, a network error is returned to Fetch. Renderer-side security checks about tainting for cross-origin content are tied to the types of filtered responses defined in Fetch.

respondWith(r) method steps are:

Let event be this.
If event’s dispatch flag is unset, throw an "InvalidStateError" DOMException.
If event’s respond-with entered flag is set, throw an "InvalidStateError" DOMException.
Add lifetime promise r to event.

Note: event.respondWith(r) extends the lifetime of the event by default as if event.waitUntil(r) is called.
Set event’s stop propagation flag and stop immediate propagation flag.
Set event’s respond-with entered flag.
Set event’s wait to respond flag.
Let targetRealm be event’s relevant Realm.
Upon rejection of r:
1. Set event’s respond-with error flag.
2. Unset event’s wait to respond flag.
Upon fulfillment of r with response:
1. If response is not a Response object, then set the respond-with error flag.
  
  Note: If the respond-with error flag is set, a network error is returned to Fetch through Handle Fetch algorithm. (See the step 21.1.) Otherwise, the value response is returned to Fetch through Handle Fetch algorithm. (See the step 22.1.)
2. Else:
  1. Let bytes be an empty byte sequence.
  2. Let end-of-body be false.
  3. Let done be false.
  4. Let potentialResponse be a copy of response’s associated response, except for its body.
  5. If response’s body is non-null, run these substeps:
    1. Let reader be the result of getting a reader from response’s body's stream.
    2. Let pullAlgorithm be an action that runs these steps:
      1. Let readRequest be a new read request with the following items:
        
        chunk steps, given chunk
        
        Assert: chunk is a Uint8Array.
        
        Append the bytes represented by chunk to bytes.
        
        Perform ! DetachArrayBuffer(chunk.[[ViewedArrayBuffer]]).
        
        close steps
        
        Set end-of-body to true.
        
        error steps
        
        error newStream with a TypeError.
      2. Read a chunk from reader given readRequest.
    3. Let cancelAlgorithm be an action that cancels reader.
    4. Let highWaterMark be a non-negative, non-NaN number, chosen by the user agent.
    5. Let sizeAlgorithm be an algorithm that accepts a chunk object and returns a non-negative, non-NaN, non-infinite number, chosen by the user agent.
    6. Let newStream be the result of creating a ReadableStream with pullAlgorithm set to pullAlgorithm, cancelAlgorithm set to cancelAlgorithm, highWaterMark set to highWaterMark, and sizeAlgorithm set to sizeAlgorithm, in targetRealm.
    7. Set potentialResponse’s body to a new body whose stream is newStream.
    8. Run these subsubsteps repeatedly in parallel while done is false:
      1. If newStream is errored, then set done to true.
      2. Otherwise, if bytes is empty and end-of-body is true, then close newStream and set done to true.
      3. Otherwise, if bytes is not empty, run these subsubsubsteps:
        
        Let chunk be a subsequence of bytes starting from the beginning of bytes.
        
        Remove chunk from bytes.
        
        Let buffer be an ArrayBuffer object created in targetRealm and containing chunk.
        
        Enqueue a Uint8Array object created in targetRealm and wrapping buffer to newStream.
    Note: These substeps are meant to produce the observable equivalent of "piping" response’s body's stream into potentialResponse.
    
    Note: The data written by the service worker in chunks are not guaranteed to read in identical chunks by the client that receives the data. That is, the client will read the same data that was written, but it may be chunked differently by the browser.
  6. Set event’s potential response to potentialResponse.
3. Unset event’s wait to respond flag.

4.6. `ExtendableMessageEvent`

ExtendableMessageEvent

In all current engines.

Firefox44+Safari11.1+Chrome51+

Opera38+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android51+Android WebView51+Samsung Internet5.0+Opera Mobile41+

ExtendableMessageEvent/ExtendableMessageEvent

In all current engines.

Firefox44+Safari11.1+Chrome51+

Opera38+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android51+Android WebView51+Samsung Internet5.0+Opera Mobile41+

[Exposed=ServiceWorker]
interface ExtendableMessageEvent : ExtendableEvent {
  constructor(DOMString type, optional ExtendableMessageEventInit eventInitDict = {});
  readonly attribute any data;
  readonly attribute USVString origin;
  readonly attribute DOMString lastEventId;
  [SameObject] readonly attribute (Client or ServiceWorker or MessagePort)? source;
  readonly attribute FrozenArray<MessagePort> ports;
};

dictionary ExtendableMessageEventInit : ExtendableEventInit {
  any data = null;
  USVString origin = "";
  DOMString lastEventId = "";
  (Client or ServiceWorker or MessagePort)? source = null;
  sequence<MessagePort> ports = [];
};

Service workers define the extendable message event to allow extending the lifetime of the event. For the message event, service workers use the ExtendableMessageEvent interface which extends the ExtendableEvent interface.

4.6.1. `event.data`

ExtendableMessageEvent/data

In all current engines.

Firefox44+Safari11.1+Chrome51+

Opera38+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android51+Android WebView51+Samsung Internet5.0+Opera Mobile41+

The data attribute must return the value it was initialized to. When the object is created, this attribute must be initialized to null. It represents the message being sent.

4.6.2. `event.origin`

ExtendableMessageEvent/origin

In all current engines.

Firefox44+Safari11.1+Chrome51+

Opera38+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android51+Android WebView51+Samsung Internet5.0+Opera Mobile41+

The origin attribute must return the value it was initialized to. When the object is created, this attribute must be initialized to the empty string. It represents the origin of the service worker client that sent the message.

4.6.3. `event.lastEventId`

ExtendableMessageEvent/lastEventId

In all current engines.

Firefox44+Safari11.1+Chrome51+

Opera38+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android51+Android WebView51+Samsung Internet5.0+Opera Mobile41+

The lastEventId attribute must return the value it was initialized to. When the object is created, this attribute must be initialized to the empty string.

4.6.4. `event.source`

ExtendableMessageEvent/source

In all current engines.

Firefox44+Safari11.1+Chrome51+

Opera38+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android51+Android WebView51+Samsung Internet5.0+Opera Mobile41+

The source attribute must return the value it was initialized to. When the object is created, this attribute must be initialized to null. It represents the Client object from which the message is sent.

4.6.5. `event.ports`

ExtendableMessageEvent/ports

In all current engines.

Firefox44+Safari11.1+Chrome51+

Opera38+Edge79+

Edge (Legacy)17+IENone

Firefox for Android44+iOS Safari11.3+Chrome for Android51+Android WebView51+Samsung Internet5.0+Opera Mobile41+

The ports attribute must return the value it was initialized to. When the object is created, this attribute must be initialized to the empty array. It represents the MessagePort array being sent.

4.7. Events

The following events, called service worker events, are dispatched on ServiceWorkerGlobalScope object:

Event name	Interface	Category	Dispatched when…
`install`	`ExtendableEvent`	Lifecycle	The service worker's containing service worker registration’s installing worker changes. (See step 11.2 of the Install algorithm.)
ServiceWorkerGlobalScope/activate_event In all current engines. Firefox44+Safari11.1+Chrome40+ Opera24+Edge79+ Edge (Legacy)17+IENone Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile24+ `activate`	`ExtendableEvent`	Lifecycle	The service worker's containing service worker registration’s active worker changes. (See step 12.2 of the Activate algorithm.)
`fetch`	`FetchEvent`	Functional	The http fetch invokes Handle Fetch with `request`. As a result of performing Handle Fetch, the service worker returns a response to the http fetch. The response, represented by a `Response` object, can be retrieved from a `Cache` object or directly from network using `self.fetch(input, init)` method. (A custom `Response` object can be another option.)
push	`PushEvent`	Functional	(See Firing a push event.)
notificationclick	`NotificationEvent`	Functional	(See Activating a notification.)
notificationclose	`NotificationEvent`	Functional	(See Closing a notification.)
sync	`SyncEvent`	Functional	(See Firing a sync event.)
canmakepayment	CanMakePaymentEvent	Functional	(See Handling a CanMakePaymentEvent.)
paymentrequest	PaymentRequestEvent	Functional	(See Handling a PaymentRequestEvent.)
ServiceWorkerGlobalScope/message_event In all current engines. Firefox44+Safari11.1+Chrome40+ Opera24+Edge79+ Edge (Legacy)17+IENone Firefox for Android44+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile24+ `message`	`ExtendableMessageEvent`	Legacy	When it receives a message.
`messageerror`	`MessageEvent`	Legacy	When it was sent a message that cannot be deserialized.

5. Caches

To allow authors to fully manage their content caches for offline use, the Window and the WorkerGlobalScope provide the asynchronous caching methods that open and manipulate Cache objects. An origin can have multiple, named Cache objects, whose contents are entirely under the control of scripts. Caches are not shared across origins, and they are completely isolated from the browser’s HTTP cache.

5.1. Constructs

A request response list is a list of pairs consisting of a request (a request) and a response (a response).

The relevant request response list is the instance that this represents.

A name to cache map is an ordered map whose entry consists of a key (a string that represents the name of a request response list) and a value (a request response list).

Each storage key has an associated name to cache map.

The relevant name to cache map is the name to cache map associated with the storage key obtained from this's associated global object's environment settings object.

5.2. Understanding Cache Lifetimes

The Cache instances are not part of the browser’s HTTP cache. The Cache objects are exactly what authors have to manage themselves. The Cache objects do not get updated unless authors explicitly request them to be. The Cache objects do not expire unless authors delete the entries. The Cache objects do not disappear just because the service worker script is updated. That is, caches are not updated automatically. Updates must be manually managed. This implies that authors should version their caches by name and make sure to use the caches only from the version of the service worker that can safely operate on.

5.3. `self.caches`

caches

In all current engines.

Firefox41+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)16+IENone

Firefox for Android41+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

partial interface mixin WindowOrWorkerGlobalScope {
  [SecureContext, SameObject] readonly attribute CacheStorage caches;
};

5.3.1. `caches`

caches getter steps are to return this's associated CacheStorage object.

5.4. `Cache`

Cache

In all current engines.

Firefox41+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)16+IENone

Firefox for Android41+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

[SecureContext, Exposed=(Window,Worker)]
interface Cache {
  [NewObject] Promise<(Response or undefined)> match(RequestInfo request, optional CacheQueryOptions options = {});
  [NewObject] Promise<FrozenArray<Response>> matchAll(optional RequestInfo request, optional CacheQueryOptions options = {});
  [NewObject] Promise<undefined> add(RequestInfo request);
  [NewObject] Promise<undefined> addAll(sequence<RequestInfo> requests);
  [NewObject] Promise<undefined> put(RequestInfo request, Response response);
  [NewObject] Promise<boolean> delete(RequestInfo request, optional CacheQueryOptions options = {});
  [NewObject] Promise<FrozenArray<Request>> keys(optional RequestInfo request, optional CacheQueryOptions options = {});
};

dictionary CacheQueryOptions {
  boolean ignoreSearch = false;
  boolean ignoreMethod = false;
  boolean ignoreVary = false;
};

A Cache object represents a request response list. Multiple separate objects implementing the Cache interface across documents and workers can all be associated with the same request response list simultaneously.

A cache batch operation is a struct that consists of:

A type ("delete" or "put").
A request (a request).
A response (a response).
An options (a CacheQueryOptions).

5.4.1. `match(request, options)`

Cache/match

In all current engines.

Firefox41+Safari11.1+Chrome43+

Opera30+Edge79+

Edge (Legacy)16+IENone

Firefox for Android41+iOS Safari11.3+Chrome for Android43+Android WebView43+Samsung Internet4.0+Opera Mobile30+

The match(request, options) method steps are:

Let promise be a new promise.
Run these substeps in parallel:
1. Let p be the result of running the algorithm specified in matchAll(request, options) method with request and options.
2. Wait until p settles.
3. If p rejects with an exception, then:
  1. Reject promise with that exception.
4. Else if p resolves with an array, responses, then:
  1. If responses is an empty array, then:
    1. Resolve promise with undefined.
  2. Else:
    1. Resolve promise with the first element of responses.
Return promise.

5.4.2. `matchAll(request, options)`

Cache/matchAll

In all current engines.

Firefox41+Safari11.1+Chrome47+

Opera34+Edge79+

Edge (Legacy)16+IENone

Firefox for Android41+iOS Safari11.3+Chrome for Android47+Android WebView47+Samsung Internet5.0+Opera Mobile34+

The matchAll(request, options) method steps are:

Let r be null.
If the optional argument request is not omitted, then:
1. If request is a Request object, then:
  1. Set r to request’s request.
  2. If r’s method is not `GET` and options.ignoreMethod is false, return a promise resolved with an empty array.
2. Else if request is a string, then:
  1. Set r to the associated request of the result of invoking the initial value of Request as constructor with request as its argument. If this throws an exception, return a promise rejected with that exception.
Let realm be this's relevant realm.
Let promise be a new promise.
Run these substeps in parallel:
1. Let responses be an empty list.
2. If the optional argument request is omitted, then:
  1. For each requestResponse of the relevant request response list:
    1. Add a copy of requestResponse’s response to responses.
3. Else:
  1. Let requestResponses be the result of running Query Cache with r and options.
  2. For each requestResponse of requestResponses:
    1. Add a copy of requestResponse’s response to responses.
4. For each response of responses:
  1. If response’s type is "opaque" and cross-origin resource policy check with promise’s relevant settings object's origin, promise’s relevant settings object, "", and response’s internal response returns blocked, then reject promise with a TypeError and abort these steps.
5. Queue a task, on promise’s relevant settings object's responsible event loop using the DOM manipulation task source, to perform the following steps:
  1. Let responseList be a list.
  2. For each response of responses:
    1. Add a new Response object associated with response and a new Headers object whose guard is "immutable" to responseList.
  3. Resolve promise with a frozen array created from responseList, in realm.
Return promise.

5.4.3. `add(request)`

Cache/add

In all current engines.

Firefox41+Safari11.1+Chrome44+

Opera31+Edge79+

Edge (Legacy)16+IENone

Firefox for Android41+iOS Safari11.3+Chrome for Android44+Android WebView44+Samsung Internet4.0+Opera Mobile32+

The add(request) method steps are:

Let requests be an array containing only request.
Let responseArrayPromise be the result of running the algorithm specified in addAll(requests) passing requests as the argument.
Return the result of reacting to responseArrayPromise with a fulfillment handler that returns undefined.

5.4.4. `addAll(requests)`

Cache/addAll

In all current engines.

Firefox41+Safari11.1+Chrome46+

Opera33+Edge79+

Edge (Legacy)16+IENone

Firefox for Android41+iOS Safari11.3+Chrome for Android46+Android WebView46+Samsung Internet5.0+Opera Mobile33+

The addAll(requests) method steps are:

Let responsePromises be an empty list.
Let requestList be an empty list.
For each request whose type is Request in requests:
1. Let r be request’s request.
2. If r’s url's scheme is not one of "http" and "https", or r’s method is not `GET`, return a promise rejected with a TypeError.
For each request in requests:
1. Let r be the associated request of the result of invoking the initial value of Request as constructor with request as its argument. If this throws an exception, return a promise rejected with that exception.
2. If r’s url's scheme is not one of "http" and "https", then:
  1. Terminate all the ongoing fetches initiated by requests with the aborted flag set.
  2. Return a promise rejected with a TypeError.
3. If r’s client's global object is a ServiceWorkerGlobalScope object, set request’s service-workers mode to "none".
4. Set r’s initiator to "fetch" and destination to "subresource".
5. Add r to requestList.
6. Let responsePromise be a new promise.
7. Run the following substeps in parallel:
  - Fetch r.
  - To process response for response, run these substeps:
    1. If response’s type is "error", or response’s status is not an ok status or is 206, reject responsePromise with a TypeError.
    2. Else if response’s header list contains a header named `Vary`, then:
      1. Let fieldValues be the list containing the elements corresponding to the field-values of the Vary header.
      2. For each fieldValue of fieldValues:
        
        If fieldValue matches "*", then:
        
        Reject responsePromise with a TypeError.
        
        Terminate all the ongoing fetches initiated by requests with the aborted flag set.
        
        Abort these steps.
  - To process response end-of-body for response, run these substeps:
    1. If response’s aborted flag is set, reject responsePromise with an "AbortError" DOMException and abort these steps.
    2. Resolve responsePromise with response.
    Note: The cache commit is allowed when the response’s body is fully received.
8. Add responsePromise to responsePromises.
Let p be the result of getting a promise to wait for all of responsePromises.
Return the result of reacting to p with a fulfillment handler that, when called with argument responses, performs the following substeps:
1. Let operations be an empty list.
2. Let index be zero.
3. For each response in responses:
  1. Let operation be a cache batch operation.
  2. Set operation’s type to "put".
  3. Set operation’s request to requestList[index].
  4. Set operation’s response to response.
  5. Append operation to operations.
  6. Increment index by one.
4. Let realm be this's relevant realm.
5. Let cacheJobPromise be a new promise.
6. Run the following substeps in parallel:
  1. Let errorData be null.
  2. Invoke Batch Cache Operations with operations. If this throws an exception, set errorData to the exception.
  3. Queue a task, on cacheJobPromise’s relevant settings object's responsible event loop using the DOM manipulation task source, to perform the following substeps:
    1. If errorData is null, resolve cacheJobPromise with undefined.
    2. Else, reject cacheJobPromise with a new exception with errorData and a user agent-defined message, in realm.
7. Return cacheJobPromise.

5.4.5. `put(request, response)`

Cache/put

In all current engines.

Firefox41+Safari11.1+Chrome43+

Opera30+Edge79+

Edge (Legacy)16+IENone

Firefox for Android41+iOS Safari11.3+Chrome for Android43+Android WebView43+Samsung Internet4.0+Opera Mobile30+

The put(request, response) method steps are:

Let innerRequest be null.
If request is a Request object, then set innerRequest to request’s request.
Else:
1. Let requestObj be the result of invoking Request's constructor with request as its argument. If this throws an exception, return a promise rejected with exception.
2. Set innerRequest to requestObj’s request.
If innerRequest’s url's scheme is not one of "http" and "https", or innerRequest’s method is not `GET`, return a promise rejected with a TypeError.
Let innerResponse be response’s response.
If innerResponse’s status is 206, return a promise rejected with a TypeError.
If innerResponse’s header list contains a header named `Vary`, then:
1. Let fieldValues be the list containing the items corresponding to the Vary header’s field-values.
2. For each fieldValue in fieldValues:
  1. If fieldValue matches "*", return a promise rejected with a TypeError.
If innerResponse’s body is disturbed or locked, return a promise rejected with a TypeError.
Let clonedResponse be a clone of innerResponse.
Let bodyReadPromise be a promise resolved with undefined.
If innerResponse’s body is non-null, run these substeps:
1. Let stream be innerResponse’s body's stream.
2. Let reader be the result of getting a reader for stream.
3. Set bodyReadPromise to the result of reading all bytes from reader.
Note: This ensures that innerResponse’s body is locked, and we have a full buffered copy of the body in clonedResponse. An implementation could optimize by streaming directly to disk rather than memory.
Let operations be an empty list.
Let operation be a cache batch operation.
Set operation’s type to "put".
Set operation’s request to innerRequest.
Set operation’s response to clonedResponse.
Append operation to operations.
Let realm be this's relevant realm.
Return the result of the fulfillment of bodyReadPromise:
1. Let cacheJobPromise be a new promise.
2. Return cacheJobPromise and run these steps in parallel:
  1. Let errorData be null.
  2. Invoke Batch Cache Operations with operations. If this throws an exception, set errorData to the exception.
  3. Queue a task, on cacheJobPromise’s relevant settings object's responsible event loop using the DOM manipulation task source, to perform the following substeps:
    1. If errorData is null, resolve cacheJobPromise with undefined.
    2. Else, reject cacheJobPromise with a new exception with errorData and a user agent-defined message, in realm.

5.4.6. `delete(request, options)`

Cache/delete

In all current engines.

Firefox41+Safari11.1+Chrome43+

Opera30+Edge79+

Edge (Legacy)16+IENone

Firefox for Android41+iOS Safari11.3+Chrome for Android43+Android WebView43+Samsung Internet4.0+Opera Mobile30+

The delete(request, options) method steps are:

Let r be null.
If request is a Request object, then:
1. Set r to request’s request.
2. If r’s method is not `GET` and options.ignoreMethod is false, return a promise resolved with false.
Else if request is a string, then:
1. Set r to the associated request of the result of invoking the initial value of Request as constructor with request as its argument. If this throws an exception, return a promise rejected with that exception.
Let operations be an empty list.
Let operation be a cache batch operation.
Set operation’s type to "delete".
Set operation’s request to r.
Set operation’s options to options.
Append operation to operations.
Let realm be this's relevant realm.
Let cacheJobPromise be a new promise.
Run the following substeps in parallel:
1. Let errorData be null.
2. Let requestResponses be the result of running Batch Cache Operations with operations. If this throws an exception, set errorData to the exception.
3. Queue a task, on cacheJobPromise’s relevant settings object's responsible event loop using the DOM manipulation task source, to perform the following substeps:
  1. If errorData is null, then:
    1. If requestResponses is not empty, resolve cacheJobPromise with true.
    2. Else, resolve cacheJobPromise with false.
  2. Else, reject cacheJobPromise with a new exception with errorData and a user agent-defined message, in realm.
Return cacheJobPromise.

5.4.7. `keys(request, options)`

Cache/keys

In all current engines.

Firefox41+Safari11.1+Chrome43+

Opera30+Edge79+

Edge (Legacy)16+IENone

Firefox for Android41+iOS Safari11.3+Chrome for Android43+Android WebView43+Samsung Internet4.0+Opera Mobile30+

The keys(request, options) method steps are:

Let r be null.
If the optional argument request is not omitted, then:
1. If request is a Request object, then:
  1. Set r to request’s request.
  2. If r’s method is not `GET` and options.ignoreMethod is false, return a promise resolved with an empty array.
2. Else if request is a string, then:
  1. Set r to the associated request of the result of invoking the initial value of Request as constructor with request as its argument. If this throws an exception, return a promise rejected with that exception.
Let realm be this's relevant realm.
Let promise be a new promise.
Run these substeps in parallel:
1. Let requests be an empty list.
2. If the optional argument request is omitted, then:
  1. For each requestResponse of the relevant request response list:
    1. Add requestResponse’s request to requests.
3. Else:
  1. Let requestResponses be the result of running Query Cache with r and options.
  2. For each requestResponse of requestResponses:
    1. Add requestResponse’s request to requests.
4. Queue a task, on promise’s relevant settings object's responsible event loop using the DOM manipulation task source, to perform the following steps:
  1. Let requestList be a list.
  2. For each request of requests:
    1. Add a new Request object associated with request and a new associated Headers object whose guard is "immutable" to requestList.
  3. Resolve promise with a frozen array created from requestList, in realm.
Return promise.

5.5. `CacheStorage`

CacheStorage

In all current engines.

Firefox41+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)16+IENone

Firefox for Android41+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

[SecureContext, Exposed=(Window,Worker)]
interface CacheStorage {
  [NewObject] Promise<(Response or undefined)> match(RequestInfo request, optional MultiCacheQueryOptions options = {});
  [NewObject] Promise<boolean> has(DOMString cacheName);
  [NewObject] Promise<Cache> open(DOMString cacheName);
  [NewObject] Promise<boolean> delete(DOMString cacheName);
  [NewObject] Promise<sequence<DOMString>> keys();
};

dictionary MultiCacheQueryOptions : CacheQueryOptions {
  DOMString cacheName;
};

Note: CacheStorage interface is designed to largely conform to ECMAScript 6 Map objects but entirely async, and with additional convenience methods. The methods, clear, forEach, entries and values, are intentionally excluded from the scope of the first version resorting to the ongoing discussion about the async iteration by TC39.

The user agent must create a CacheStorage object when a Window object or a WorkerGlobalScope object is created and associate it with that global object.

A CacheStorage object represents a name to cache map of its associated global object's environment settings object’s origin. Multiple separate objects implementing the CacheStorage interface across documents and workers can all be associated with the same name to cache map simultaneously.

5.5.1. `match(request, options)`

CacheStorage/match

In all current engines.

Firefox41+Safari11.1+Chrome54+

Opera41+Edge79+

Edge (Legacy)16+IENone

Firefox for Android41+iOS Safari11.3+Chrome for Android54+Android WebView54+Samsung Internet6.0+Opera Mobile41+

The match(request, options) method steps are:

If options["cacheName"] exists, then:
1. Return a new promise promise and run the following substeps in parallel:
  1. For each cacheName → cache of the relevant name to cache map:
    1. If options["cacheName"] matches cacheName, then:
      1. Resolve promise with the result of running the algorithm specified in match(request, options) method of Cache interface with request and options (providing cache as thisArgument to the [[Call]] internal method of match(request, options).)
      2. Abort these steps.
  2. Resolve promise with undefined.
Else:
1. Let promise be a promise resolved with undefined.
2. For each cacheName → cache of the relevant name to cache map:
  1. Set promise to the result of reacting to itself with a fulfillment handler that, when called with argument response, performs the following substeps:
    1. If response is not undefined, return response.
    2. Return the result of running the algorithm specified in match(request, options) method of Cache interface with request and options as the arguments (providing cache as thisArgument to the [[Call]] internal method of match(request, options).)
3. Return promise.

5.5.2. `has(cacheName)`

CacheStorage/has

In all current engines.

Firefox41+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)16+IENone

Firefox for Android41+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

The has(cacheName) method steps are:

Let promise be a new promise.
Run the following substeps in parallel:
1. For each key → value of the relevant name to cache map:
  1. If cacheName matches key, resolve promise with true and abort these steps.
2. Resolve promise with false.
Return promise.

5.5.3. `open(cacheName)`

CacheStorage/open

In all current engines.

Firefox41+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)16+IENone

Firefox for Android41+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

The open(cacheName) method steps are:

Let promise be a new promise.
Run the following substeps in parallel:
1. For each key → value of the relevant name to cache map:
  1. If cacheName matches key, then:
    1. Resolve promise with a new Cache object that represents value.
    2. Abort these steps.
2. Let cache be a new request response list.
3. Set the relevant name to cache map[cacheName] to cache. If this cache write operation failed due to exceeding the granted quota limit, reject promise with a "QuotaExceededError" DOMException and abort these steps.
4. Resolve promise with a new Cache object that represents cache.
Return promise.

5.5.4. `delete(cacheName)`

CacheStorage/delete

In all current engines.

Firefox41+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)16+IENone

Firefox for Android41+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

The delete(cacheName) method steps are:

Let promise be the result of running the algorithm specified in has(cacheName) method with cacheName.
Return the result of reacting to promise with a fulfillment handler that, when called with argument cacheExists, performs the following substeps:
1. If cacheExists is false, then:
  1. Return false.
2. Let cacheJobPromise be a new promise.
3. Run the following substeps in parallel:
  1. Remove the relevant name to cache map[cacheName].
  2. Resolve cacheJobPromise with true.
  Note: After this step, the existing DOM objects (i.e. the currently referenced Cache, Request, and Response objects) should remain functional.
4. Return cacheJobPromise.

5.5.5. `keys()`

CacheStorage/keys

In all current engines.

Firefox41+Safari11.1+Chrome40+

Opera27+Edge79+

Edge (Legacy)16+IENone

Firefox for Android41+iOS Safari11.3+Chrome for Android40+Android WebView40+Samsung Internet4.0+Opera Mobile27+

The keys() method steps are:

Let promise be a new promise.
Run the following substeps in parallel:
1. Let cacheKeys be the result of getting the keys of the relevant name to cache map.
  
  Note: The items in the result ordered set are in the order that their corresponding entry was added to the name to cache map.
2. Resolve promise with cacheKeys.
Return promise.

6. Security Considerations

6.1. Secure Context

Service workers must execute in secure contexts. Service worker clients must also be secure contexts to register a service worker registration, to get access to the service worker registrations and the service workers, to do messaging with the service workers, and to be manipulated by the service workers.

Note: This effectively means that service workers and their service worker clients need to be hosted over HTTPS. A user agent can allow localhost (see the requirements), 127.0.0.0/8, and ::1/128 for development purposes. The primary reason for this restriction is to protect users from the risks associated with insecure contexts.

6.2. Content Security Policy

Whenever a user agent invokes the Run Service Worker algorithm with a service worker serviceWorker:

If serviceWorker’s script resource was delivered with a Content-Security-Policy HTTP header containing the value policy, the user agent must enforce policy for serviceWorker.
If serviceWorker’s script resource was delivered with a Content-Security-Policy-Report-Only HTTP header containing the value policy, the user agent must monitor policy for serviceWorker.

The primary reason for this restriction is to mitigate a broad class of content injection vulnerabilities, such as cross-site scripting (XSS).

6.3. Origin Relativity

6.3.1. Origin restriction

This section is non-normative.

A service worker executes in the registering service worker client's origin. One of the advanced concerns that major applications would encounter is whether they can be hosted from a CDN. By definition, these are servers in other places, often on other origins. Therefore, service workers cannot be hosted on CDNs. But they can include resources via importScripts(). The reason for this restriction is that service workers create the opportunity for a bad actor to turn a bad day into a bad eternity.

6.3.2. `importScripts(urls)`

When the importScripts(urls) method is called on a ServiceWorkerGlobalScope object, the user agent must import scripts into worker global scope, given this ServiceWorkerGlobalScope object and urls, and with the following steps to perform the fetch given the request request:

Let serviceWorker be request’s client's global object's service worker.
Let map be serviceWorker’s script resource map.
Let url be request’s url.
If serviceWorker’s state is not "parsed" or "installing":
1. Return map[url] if it exists and a network error otherwise.
If map[url] exists:
1. Append url to serviceWorker’s set of used scripts.
2. Return map[url].
Let registration be serviceWorker’s containing service worker registration.
Set request’s service-workers mode to "none".
Set request’s cache mode to "no-cache" if any of the following are true:
- registration’s update via cache mode is "none".
- The current global object's force bypass cache for import scripts flag is set.
- registration is stale.
Let response be the result of fetching request.
If response’s cache state is not "local", set registration’s last update check time to the current time.
If response’s unsafe response is a bad import script response, then return a network error.
Set map[url] to response.
Append url to serviceWorker’s set of used scripts.
Set serviceWorker’s classic scripts imported flag.
Return response.

6.4. Cross-Origin Resources and CORS

This section is non-normative.

Applications tend to cache items that come from a CDN or other origin. It is possible to request many of them directly using <script>, <img>, <video> and <link> elements. It would be hugely limiting if this sort of runtime collaboration broke when offline. Similarly, it is possible to fetch many sorts of off-origin resources when appropriate CORS headers are set. Service workers enable this by allowing Caches to fetch and cache off-origin items. Some restrictions apply, however. First, unlike same-origin resources which are managed in the Cache as Response objects whose corresponding responses are basic filtered response, the objects stored are Response objects whose corresponding responses are either CORS filtered responses or opaque filtered responses. They can be passed to event.respondWith(r) method in the same manner as the Response objects whose corresponding responses are basic filtered responses, but cannot be meaningfully created programmatically. These limitations are necessary to preserve the security invariants of the platform. Allowing Caches to store them allows applications to avoid re-architecting in most cases.

6.5. Path restriction

This section is non-normative.

In addition to the origin restriction, service workers are restricted by the path of the service worker script. For example, a service worker script at https://www.example.com/~bob/sw.js can be registered for the scope url https://www.example.com/~bob/ but not for the scope https://www.example.com/ or https://www.example.com/~alice/. This provides some protection for sites that host multiple-user content in separated directories on the same origin. However, the path restriction is not considered a hard security boundary, as only origins are. Sites are encouraged to use different origins to securely isolate segments of the site if appropriate.

Servers can remove the path restriction by setting a Service-Worker-Allowed header on the service worker script.

6.6. Service worker script request

This section is non-normative.

To further defend against malicious registration of a service worker on a site, this specification requires that:

The Service-Worker header is present on service worker script requests, and
Service worker scripts are served with a JavaScript MIME type.

6.7. Implementer Concerns

This section is non-normative.

The implementers are encouraged to note:

Plug-ins should not load via service workers. As plug-ins may get their security origins from their own urls, the embedding service worker cannot handle it. For this reason, the Handle Fetch algorithm makes <embed> and <object> requests immediately fallback to the network without dispatching fetch event.
Some of the legacy networking stack code may need to be carefully audited to understand the ramifications of interactions with service workers.

6.8. Privacy

Service workers introduce new persistent storage features including registration map (for service worker registrations and their service workers), request response list and name to cache map (for caches), and script resource map (for script resources). In order to protect users from any potential unsanctioned tracking threat, these persistent storages should be cleared when users intend to clear them and should maintain and interoperate with existing user controls e.g. purging all existing persistent storages.

7. Extensibility

Service Workers specification is extensible from other specifications.

7.1. Define API bound to Service Worker Registration

Specifications may define an API tied to a service worker registration by using partial interface definition to the ServiceWorkerRegistration interface where it may define the specification specific attributes and methods:

partial interface ServiceWorkerRegistration {
  // e.g. define an API namespace
  readonly attribute APISpaceType APISpace;
  // e.g. define a method
  Promise<T> methodName(/* list of arguments */);
};

7.2. Define Functional Event

Specifications may define a functional event by extending ExtendableEvent interface:

// e.g. define FunctionalEvent interface
interface FunctionalEvent : ExtendableEvent {
  // add a functional event’s own attributes and methods
};

7.3. Define Event Handler

Specifications may define an event handler attribute for the corresponding functional event using partial interface definition to the ServiceWorkerGlobalScope interface:

partial interface ServiceWorkerGlobalScope {
  attribute EventHandler onfunctionalevent;
};

7.4. Firing Functional Events

To request a functional event dispatch to the active worker of a service worker registration, specifications should invoke Fire Functional Event.

Appendix A: Algorithms

The following definitions are the user agent’s internal data structures used throughout the specification.

A registration map is an ordered map where the keys are (storage key, serialized scope urls) and the values are service worker registrations.

A job is an abstraction of one of register, update, and unregister request for a service worker registration.

A job has a job type, which is one of register, update, and unregister.

A job has a storage key (a storage key).

A job has a scope url (a URL).

A job has a script url (a URL).

A job has a worker type ("classic" or "module").

A job has an update via cache mode, which is "imports", "all", or "none".

A job has a client (a service worker client). It is initially null.

A job has a referrer (a URL or null).

A job has a job promise (a promise). It is initially null.

A job has a containing job queue (a job queue or null). It is initially null.

A job has a list of equivalent jobs (a list of jobs). It is initially the empty list.

A job has a force bypass cache flag. It is initially unset.

Two jobs are equivalent when their job type is the same and:

For register and update jobs, their scope url, script url, worker type, and update via cache mode are the same.
For unregister jobs, their scope url is the same.

A job queue is a thread safe queue used to synchronize the set of concurrent jobs. The job queue contains jobs as its items. A job queue is initially empty.

A scope to job queue map is an ordered map where the keys are scope urls, serialized, and the values are job queues.

A bad import script response is a response for which any of the following conditions are met:

response’s type is "error"
response’s status is not an ok status
The result of extracting a MIME type from response’s header list is not a JavaScript MIME type

Note: Keep this definition in sync with fetch a classic worker-imported script.

Create Job

Input: jobType, a job type; storage key, a storage key; scopeURL, a URL; scriptURL, a URL; promise, a promise; client, a service worker client
Output: job, a job

Let job be a new job.
Set job’s job type to jobType.
Set job’s storage key to storage key.
Set job’s scope url to scopeURL.
Set job’s script url to scriptURL.
Set job’s job promise to promise.
Set job’s client to client.
If client is not null, set job’s referrer to client’s creation URL.
Return job.

Schedule Job

Input: job, a job
Output: none

Let jobQueue be null.
Let jobScope be job’s scope url, serialized.
If scope to job queue map[jobScope] does not exist, set scope to job queue map[jobScope] to a new job queue.
Set jobQueue to scope to job queue map[jobScope].
If jobQueue is empty, then:
1. Set job’s containing job queue to jobQueue, and enqueue job to jobQueue.
2. Invoke Run Job with jobQueue.
Else:
1. Let lastJob be the element at the back of jobQueue.
2. If job is equivalent to lastJob and lastJob’s job promise has not settled, append job to lastJob’s list of equivalent jobs.
3. Else, set job’s containing job queue to jobQueue, and enqueue job to jobQueue.

Run Job

Input: jobQueue, a job queue
Output: none

Assert: jobQueue is not empty.
Queue a task to run these steps:
1. Let job be the first item in jobQueue.
2. If job’s job type is register, run Register with job in parallel.
3. Else if job’s job type is update, run Update with job in parallel.
  
  Note: For a register job and an update job, the user agent delays queuing a task for running the job until after a DOMContentLoaded event has been dispatched to the document that initiated the job.
4. Else if job’s job type is unregister, run Unregister with job in parallel.

Finish Job

Input: job, a job
Output: none

Let jobQueue be job’s containing job queue.
Assert: the first item in jobQueue is job.
Dequeue from jobQueue.
If jobQueue is not empty, invoke Run Job with jobQueue.

Resolve Job Promise

Input: job, a job; value, any
Output: none

If job’s client is not null, queue a task, on job’s client's responsible event loop using the DOM manipulation task source, to run the following substeps:
1. Let convertedValue be null.
2. If job’s job type is either register or update, set convertedValue to the result of getting the service worker registration object that represents value in job’s client.
3. Else, set convertedValue to value, in job’s client's Realm.
4. Resolve job’s job promise with convertedValue.
For each equivalentJob in job’s list of equivalent jobs:
1. If equivalentJob’s client is null, continue to the next iteration of the loop.
2. Queue a task, on equivalentJob’s client's responsible event loop using the DOM manipulation task source, to run the following substeps:
  1. Let convertedValue be null.
  2. If equivalentJob’s job type is either register or update, set convertedValue to the result of getting the service worker registration object that represents value in equivalentJob’s client.
  3. Else, set convertedValue to value, in equivalentJob’s client's Realm.
  4. Resolve equivalentJob’s job promise with convertedValue.

Reject Job Promise

Input: job, a job; errorData, the information necessary to create an exception
Output: none

If job’s client is not null, queue a task, on job’s client's responsible event loop using the DOM manipulation task source, to reject job’s job promise with a new exception with errorData and a user agent-defined message, in job’s client's Realm.
For each equivalentJob in job’s list of equivalent jobs:
1. If equivalentJob’s client is null, continue.
2. Queue a task, on equivalentJob’s client's responsible event loop using the DOM manipulation task source, to reject equivalentJob’s job promise with a new exception with errorData and a user agent-defined message, in equivalentJob’s client's Realm.

Start Register

Input: scopeURL, a URL or failure or null; scriptURL, a URL or failure; promise, a promise; client, a service worker client; referrer, a URL; workerType, a worker type; updateViaCache, an update via cache mode
Output: none

If scriptURL is failure, reject promise with a TypeError and abort these steps.
Set scriptURL’s fragment to null.

Note: The user agent does not store the fragment of the script’s url. This means that the fragment does not have an effect on identifying service workers.
If scriptURL’s scheme is not one of "http" and "https", reject promise with a TypeError and abort these steps.
If any of the strings in scriptURL’s path contains either ASCII case-insensitive "%2f" or ASCII case-insensitive "%5c", reject promise with a TypeError and abort these steps.
If scopeURL is null, set scopeURL to the result of parsing the string "./" with scriptURL.

Note: The scope url for the registration is set to the location of the service worker script by default.
If scopeURL is failure, reject promise with a TypeError and abort these steps.
Set scopeURL’s fragment to null.

Note: The user agent does not store the fragment of the scope url. This means that the fragment does not have an effect on identifying service worker registrations.
If scopeURL’s scheme is not one of "http" and "https", reject promise with a TypeError and abort these steps.
If any of the strings in scopeURL’s path contains either ASCII case-insensitive "%2f" or ASCII case-insensitive "%5c", reject promise with a TypeError and abort these steps.
Let storage key be the result of running obtain a storage key given client.
Let job be the result of running Create Job with register, storage key, scopeURL, scriptURL, promise, and client.
Set job’s worker type to workerType.
Set job’s update via cache mode to updateViaCache.
Set job’s referrer to referrer.
Invoke Schedule Job with job.

Register

Input: job, a job
Output: none

If the result of running potentially trustworthy origin with the origin of job’s script url as the argument is Not Trusted, then:
1. Invoke Reject Job Promise with job and "SecurityError" DOMException.
2. Invoke Finish Job with job and abort these steps.
If job’s script url's origin and job’s referrer's origin are not same origin, then:
1. Invoke Reject Job Promise with job and "SecurityError" DOMException.
2. Invoke Finish Job with job and abort these steps.
If job’s scope url's origin and job’s referrer's origin are not same origin, then:
1. Invoke Reject Job Promise with job and "SecurityError" DOMException.
2. Invoke Finish Job with job and abort these steps.
Let registration be the result of running Get Registration given job’s storage key and job’s scope url.
If registration is not null, then:
1. Let newestWorker be the result of running the Get Newest Worker algorithm passing registration as the argument.
2. If newestWorker is not null, job’s script url equals newestWorker’s script url, job’s worker type equals newestWorker’s type, and job’s update via cache mode's value equals registration’s update via cache mode, then:
  1. Invoke Resolve Job Promise with job and registration.
  2. Invoke Finish Job with job and abort these steps.
Else:
1. Invoke Set Registration algorithm with job’s storage key, job’s scope url, and job’s update via cache mode.
Invoke Update algorithm passing job as the argument.

Update

Input: job, a job
Output: none

Let registration be the result of running Get Registration given job’s storage key and job’s scope url.
If registration is null, then:
1. Invoke Reject Job Promise with job and TypeError.
2. Invoke Finish Job with job and abort these steps.
Let newestWorker be the result of running Get Newest Worker algorithm passing registration as the argument.
If job’s job type is update, and newestWorker is not null and its script url does not equal job’s script url, then:
1. Invoke Reject Job Promise with job and TypeError.
2. Invoke Finish Job with job and abort these steps.
Let referrerPolicy be the empty string.
Let hasUpdatedResources be false.
Let updatedResourceMap be an ordered map where the keys are URLs and the values are responses.
Switching on job’s worker type, run these substeps with the following options:

"classic"

Fetch a classic worker script given job’s serialized script url, job’s client, "serviceworker", and the to-be-created environment settings object for this service worker.

"module"

Fetch a module worker script graph given job’s serialized script url, job’s client, "serviceworker", "omit", and the to-be-created environment settings object for this service worker.

Using the to-be-created environment settings object rather than a concrete environment settings object. This is used due to the unique processing model of service workers compared to the processing model of other web workers. The script fetching algorithms of HTML standard originally designed for other web workers require an environment settings object of the execution environment, but service workers fetch a script separately in the Update algorithm before the script later runs multiple times through the Run Service Worker algorithm.

The fetch a classic worker script algorithm and the fetch a module worker script graph algorithm in HTML take job’s client as an argument. job’s client is null when passed from the Soft Update algorithm.

To perform the fetch given request, run the following steps:
1. Append `Service-Worker`/`script` to request’s header list.
  
  Note: See the definition of the Service-Worker header in Appendix B: Extended HTTP headers.
2. Set request’s cache mode to "no-cache" if any of the following are true:
  - registration’s update via cache mode is not "all".
  - job’s force bypass cache flag is set.
  - newestWorker is not null and registration is stale.
  Note: Even if the cache mode is not set to "no-cache", the user agent obeys Cache-Control header’s max-age value in the network layer to determine if it should bypass the browser cache.
3. Set request’s service-workers mode to "none".
4. If the is top-level flag is unset, then return the result of fetching request.
5. Set request’s redirect mode to "error".
6. Fetch request, and asynchronously wait to run the remaining steps as part of fetch’s process response for the response response.
7. Extract a MIME type from the response’s header list. If this MIME type (ignoring parameters) is not a JavaScript MIME type, then:
  1. Invoke Reject Job Promise with job and "SecurityError" DOMException.
  2. Asynchronously complete these steps with a network error.
8. Let serviceWorkerAllowed be the result of extracting header list values given `Service-Worker-Allowed` and response’s header list.
  
  Note: See the definition of the Service-Worker-Allowed header in Appendix B: Extended HTTP headers.
9. Set policyContainer to the result of creating a policy container from a fetch response given response.
10. If serviceWorkerAllowed is failure, then:
  1. Asynchronously complete these steps with a network error.
11. Let scopeURL be registration’s scope url.
12. Let maxScopeString be null.
13. If serviceWorkerAllowed is null, then:
  1. Let resolvedScope be the result of parsing "./" using job’s script url as the base URL.
  2. Set maxScopeString to "/", followed by the strings in resolvedScope’s path (including empty strings), separated from each other by "/".
    
    Note: The final item in resolvedScope’s path will always be an empty string, so maxScopeString will have a trailing "/".
14. Else:
  1. Let maxScope be the result of parsing serviceWorkerAllowed using job’s script url as the base URL.
  2. If maxScope’s origin is job’s script url's origin, then:
    1. Set maxScopeString to "/", followed by the strings in maxScope’s path (including empty strings), separated from each other by "/".
15. Let scopeString be "/", followed by the strings in scopeURL’s path (including empty strings), separated from each other by "/".
16. If maxScopeString is null or scopeString does not start with maxScopeString, then:
  1. Invoke Reject Job Promise with job and "SecurityError" DOMException.
  2. Asynchronously complete these steps with a network error.
17. Let url be request’s url.
18. Set updatedResourceMap[url] to response.
19. If response’s cache state is not "local", set registration’s last update check time to the current time.
20. Set hasUpdatedResources to true if any of the following are true:
  - newestWorker is null.
  - newestWorker’s script url is not url or newestWorker’s type is not job’s worker type.
  - newestWorker’s script resource map[url]'s body is not byte-for-byte identical with response’s body.
21. If hasUpdatedResources is false and newestWorker’s classic scripts imported flag is set, then:
  
  Note: The following checks to see if an imported script has been updated, since the main script has not changed.
  1. For each importUrl → storedResponse of newestWorker’s script resource map:
    1. If importUrl is url, then continue.
    2. Let importRequest be a new request whose url is importUrl, client is job’s client, destination is "script", parser metadata is "not parser-inserted", synchronous flag is set, and whose use-URL-credentials flag is set.
    3. Set importRequest’s cache mode to "no-cache" if any of the following are true:
      - registration’s update via cache mode is "none".
      - job’s force bypass cache flag is set.
      - registration is stale.
    4. Let fetchedResponse be the result of fetching importRequest.
    5. Set updatedResourceMap[importRequest’s url] to fetchedResponse.
    6. Set fetchedResponse to fetchedResponse’s unsafe response.
    7. If fetchedResponse’s cache state is not "local", set registration’s last update check time to the current time.
    8. If fetchedResponse is a bad import script response, continue.
      
      Note: Bad responses for importScripts() are ignored for the purpose of the byte-to-byte check. Only good responses for the incumbent worker and good responses for the potential update worker are considered. See issue #1374 for some rationale.
    9. If fetchedResponse’s body is not byte-for-byte identical with storedResponse’s unsafe response's body, set hasUpdatedResources to true.
      
      Note: The control does not break the loop in this step to continue with all the imported scripts to populate the cache.
22. Asynchronously complete these steps with response.
When the algorithm asynchronously completes, continue the rest of these steps, with script being the asynchronous completion value.
If script is null or Is Async Module with script’s record, script’s base URL, and « » is true, then:
1. Invoke Reject Job Promise with job and TypeError.
  
  Note: This will do nothing if Reject Job Promise was previously invoked with "SecurityError" DOMException.
2. If newestWorker is null, then remove registration map[(registration’s storage key, serialized scopeURL)].
3. Invoke Finish Job with job and abort these steps.
If hasUpdatedResources is false, then:
1. Set registration’s update via cache mode to job’s update via cache mode.
2. Invoke Resolve Job Promise with job and registration.
3. Invoke Finish Job with job and abort these steps.
Let worker be a new service worker.
Set worker’s script url to job’s script url, worker’s script resource to script, worker’s type to job’s worker type, and worker’s script resource map to updatedResourceMap.
Append url to worker’s set of used scripts.
Set worker’s script resource’s policy container to policyContainer.
Let forceBypassCache be true if job’s force bypass cache flag is set, and false otherwise.
Let runResult be the result of running the Run Service Worker algorithm with worker and forceBypassCache.
If runResult is failure or an abrupt completion, then:
1. Invoke Reject Job Promise with job and TypeError.
2. If newestWorker is null, then remove registration map[(registration’s storage key, serialized scopeURL)].
3. Invoke Finish Job with job.
Else, invoke Install algorithm with job, worker, and registration as its arguments.

Soft Update

The user agent may call this as often as it likes to check for updates.

Input

registration, a service worker registration

forceBypassCache, an optional boolean, false by default

Note: Implementers may use forceBypassCache to aid debugging (e.g. invocations from developer tools), and other specifications that extend service workers may also use the flag on their own needs.

Output

None

Let newestWorker be the result of running Get Newest Worker algorithm passing registration as its argument.
If newestWorker is null, abort these steps.
Let job be the result of running Create Job with update, registration’s storage key, registration’s scope url, newestWorker’s script url, null, and null.
Set job’s worker type to newestWorker’s type.
Set job’s force bypass cache flag if forceBypassCache is true.
Invoke Schedule Job with job.

Install

Input: job, a job; worker, a service worker; registration, a service worker registration
Output: none

Let installFailed be false.
Let newestWorker be the result of running Get Newest Worker algorithm passing registration as its argument.
Set registration’s update via cache mode to job’s update via cache mode.
Run the Update Registration State algorithm passing registration, "installing" and worker as the arguments.
Run the Update Worker State algorithm passing registration’s installing worker and "installing" as the arguments.
Assert: job’s job promise is not null.
Invoke Resolve Job Promise with job and registration.
Let settingsObjects be all environment settings objects whose origin is registration’s scope url's origin.
For each settingsObject of settingsObjects, queue a task on settingsObject’s responsible event loop in the DOM manipulation task source to run the following steps:
1. Let registrationObjects be every ServiceWorkerRegistration object in settingsObject’s realm, whose service worker registration is registration.
2. For each registrationObject of registrationObjects, fire an event on registrationObject named updatefound.
Let installingWorker be registration’s installing worker.
If the result of running the Should Skip Event algorithm with installingWorker and "install" is false, then:
1. Let forceBypassCache be true if job’s force bypass cache flag is set, and false otherwise.
2. If the result of running the Run Service Worker algorithm with installingWorker and forceBypassCache is failure, then:
  1. Set installFailed to true.
3. Else:
  1. Queue a task task on installingWorker’s event loop using the DOM manipulation task source to run the following steps:
    1. Let e be the result of creating an event with ExtendableEvent.
    2. Initialize e’s type attribute to install.
    3. Dispatch e at installingWorker’s global object.
    4. WaitForAsynchronousExtensions: Run the following substeps in parallel:
      1. Wait until e is not active.
      2. If e’s timed out flag is set, set installFailed to true.
      3. Let p be the result of getting a promise to wait for all of e’s extend lifetime promises.
      4. Upon rejection of p, set installFailed to true.
    If task is discarded, set installFailed to true.
  2. Wait for task to have executed or been discarded.
  3. Wait for the step labeled WaitForAsynchronousExtensions to complete.
If installFailed is true, then:
1. Run the Update Worker State algorithm passing registration’s installing worker and "redundant" as the arguments.
2. Run the Update Registration State algorithm passing registration, "installing" and null as the arguments.
3. If newestWorker is null, then remove registration map[(registration’s storage key, serialized registration’s scope url)].
4. Invoke Finish Job with job and abort these steps.
Let map be registration’s installing worker's script resource map.
Let usedSet be registration’s installing worker's set of used scripts.
For each url of map:
1. If usedSet does not contain url, then remove map[url].
If registration’s waiting worker is not null, then:
1. Terminate registration’s waiting worker.
2. Run the Update Worker State algorithm passing registration’s waiting worker and "redundant" as the arguments.
Run the Update Registration State algorithm passing registration, "waiting" and registration’s installing worker as the arguments.
Run the Update Registration State algorithm passing registration, "installing" and null as the arguments.
Run the Update Worker State algorithm passing registration’s waiting worker and "installed" as the arguments.
Invoke Finish Job with job.
Wait for all the tasks queued by Update Worker State invoked in this algorithm to have executed.
Invoke Try Activate with registration.

Note: If Try Activate does not trigger Activate here, Activate is tried again when the last client controlled by the existing active worker is unloaded, skipWaiting() is asynchronously called, or the extend lifetime promises for the existing active worker settle.

Activate

Input: registration, a service worker registration
Output: None

If registration’s waiting worker is null, abort these steps.
If registration’s active worker is not null, then:
1. Terminate registration’s active worker.
2. Run the Update Worker State algorithm passing registration’s active worker and "redundant" as the arguments.
Run the Update Registration State algorithm passing registration, "active" and registration’s waiting worker as the arguments.
Run the Update Registration State algorithm passing registration, "waiting" and null as the arguments.
Run the Update Worker State algorithm passing registration’s active worker and "activating" as the arguments.

Note: Once an active worker is activating, neither a runtime script error nor a force termination of the active worker prevents the active worker from getting activated.

Note: Make sure to design activation handlers to do non-essential work (like cleanup). This is because activation handlers may not all run to completion, especially in the case of browser termination during activation. A Service Worker should be designed to function properly, even if the activation handlers do not all complete successfully.
Let matchedClients be a list of service worker clients whose creation URL matches registration’s storage key and registration’s scope url.
For each client of matchedClients, queue a task on client’s responsible event loop, using the DOM manipulation task source, to run the following substeps:
1. Let readyPromise be client’s global object's ServiceWorkerContainer object’s ready promise.
2. If readyPromise is null, then continue.
3. If readyPromise is pending, resolve readyPromise with the the result of getting the service worker registration object that represents registration in readyPromise’s relevant settings object.
For each client of matchedClients:
1. If client is a window client, unassociate client’s responsible document from its application cache, if it has one.
2. Else if client is a shared worker client, unassociate client’s global object from its application cache, if it has one.
Note: Resources will now use the service worker registration instead of the existing application cache.
For each service worker client client who is using registration:
1. Set client’s active worker to registration’s active worker.
2. Invoke Notify Controller Change algorithm with client as the argument.
Let activeWorker be registration’s active worker.
If the result of running the Should Skip Event algorithm with activeWorker and "activate" is false, then:
1. If the result of running the Run Service Worker algorithm with activeWorker is not failure, then:
  1. Queue a task task on activeWorker’s event loop using the DOM manipulation task source to run the following steps:
    1. Let e be the result of creating an event with ExtendableEvent.
    2. Initialize e’s type attribute to activate.
    3. Dispatch e at activeWorker’s global object.
    4. WaitForAsynchronousExtensions: Wait, in parallel, until e is not active.
  2. Wait for task to have executed or been discarded.
  3. Wait for the step labeled WaitForAsynchronousExtensions to complete.
Run the Update Worker State algorithm passing registration’s active worker and "activated" as the arguments.

Try Activate

Input: registration, a service worker registration
Output: None

If registration’s waiting worker is null, return.
If registration’s active worker is not null and registration’s active worker's state is "activating", return.

Note: If the existing active worker is still in activating state, the activation of the waiting worker is delayed.
Invoke Activate with registration if either of the following is true:
- registration’s active worker is null.
- The result of running Service Worker Has No Pending Events with registration’s active worker is true, and no service worker client is using registration or registration’s waiting worker's skip waiting flag is set.

Run Service Worker

Input: serviceWorker, a service worker; forceBypassCache, an optional boolean, false by default
Output: a Completion or failure

Note: This algorithm blocks until the service worker is running or fails to start.

Let unsafeCreationTime be the unsafe shared current time.
If serviceWorker is running, then return serviceWorker’s start status.
If serviceWorker’s state is "redundant", then return failure.
Assert: serviceWorker’s start status is null.
Let script be serviceWorker’s script resource.
Assert: script is not null.
Let startFailed be false.
Let agent be the result of obtaining a service worker agent, and run the following steps in that context:
1. Let realmExecutionContext be the result of creating a new JavaScript realm given agent and the following customizations:
  - For the global object, create a new ServiceWorkerGlobalScope object. Let workerGlobalScope be the created object.
2. Set serviceWorker’s global object to workerGlobalScope.
3. Let settingsObject be a new environment settings object whose algorithms are defined as follows:
  
  The realm execution context
  
  Return realmExecutionContext.
  
  The module map
  
  Return workerGlobalScope’s module map.
  
  The API URL character encoding
  
  Return UTF-8.
  
  The API base URL
  
  Return serviceWorker’s script url.
  
  The origin
  
  Return its registering service worker client's origin.
  
  The policy container
  
  Return workerGlobalScope’s policy container.
  
  The time origin
  
  Return the result of coarsening unsafeCreationTime given workerGlobalScope’s cross-origin isolated capability.
4. Set settingsObject’s id to a new unique opaque string, creation URL to serviceWorker’s script url, top-level creation URL to null, top-level origin to an implementation-defined value, target browsing context to null, and active service worker to null.
5. Set workerGlobalScope’s url to serviceWorker’s script url.
6. Set workerGlobalScope’s policy container to serviceWorker’s script resource’s policy container.
7. Set workerGlobalScope’s type to serviceWorker’s type.
8. Set workerGlobalScope’s force bypass cache for import scripts flag if forceBypassCache is true.
9. Create a new WorkerLocation object and associate it with workerGlobalScope.
10. If the run CSP initialization for a global object algorithm returns "Blocked" when executed upon workerGlobalScope, set startFailed to true and abort these steps.
11. If serviceWorker is an active worker, and there are any tasks queued in serviceWorker’s containing service worker registration’s task queues, queue them to serviceWorker’s event loop’s task queues in the same order using their original task sources.
12. Let evaluationStatus be null.
13. If script is a classic script, then:
  1. Set evaluationStatus to the result of running the classic script script.
  2. If evaluationStatus.[[Value]] is empty, this means the script was not evaluated. Set startFailed to true and abort these steps.
14. Otherwise, if script is a module script, then:
  1. Let evaluationPromise be the result of running the module script script, with report errors set to false.
  2. Assert: evaluationPromise.[[PromiseState]] is not "pending".
  3. If evaluationPromise.[[PromiseState]] is "rejected":
    1. Set evaluationStatus to ThrowCompletion(evaluationPromise.[[PromiseResult]]).
  4. Otherwise:
    1. Set evaluationStatus to NormalCompletion(undefined).
15. If the script was aborted by the Terminate Service Worker algorithm, set startFailed to true and abort these steps.
16. Set serviceWorker’s start status to evaluationStatus.
17. If script’s has ever been evaluated flag is unset, then:
  1. For each eventType of settingsObject’s global object's associated list of event listeners' event types:
    1. Append eventType to workerGlobalScope’s associated service worker's set of event types to handle.
    Note: If the global object’s associated list of event listeners does not have any event listener added at this moment, the service worker’s set of event types to handle remains an empty set. The user agents are encouraged to show a warning that the event listeners must be added on the very first evaluation of the worker script.
  2. Set script’s has ever been evaluated flag.
18. Run the responsible event loop specified by settingsObject until it is destroyed.
19. Empty workerGlobalScope’s list of active timers.
Wait for serviceWorker to be running, or for startFailed to be true.
If startFailed is true, then return failure.
Return serviceWorker’s start status.

Terminate Service Worker

Input: serviceWorker, a service worker
Output: None

Run the following steps in parallel with serviceWorker’s main loop:
1. Let serviceWorkerGlobalScope be serviceWorker’s global object.
2. Set serviceWorkerGlobalScope’s closing flag to true.
3. Remove all the items from serviceWorker’s set of extended events.
4. If there are any tasks, whose task source is either the handle fetch task source or the handle functional event task source, queued in serviceWorkerGlobalScope’s event loop’s task queues, queue them to serviceWorker’s containing service worker registration’s corresponding task queues in the same order using their original task sources, and discard all the tasks (including tasks whose task source is neither the handle fetch task source nor the handle functional event task source) from serviceWorkerGlobalScope’s event loop’s task queues without processing them.
  
  Note: This effectively means that the fetch events and the other functional events such as push events are backed up by the registration’s task queues while the other tasks including message events are discarded.
5. Abort the script currently running in serviceWorker.
6. Set serviceWorker’s start status to null.

Handle Fetch

Headers/Service-Worker-Navigation-Preload

In all current engines.

Firefoxpreview+Safari15.4+Chrome59+

Opera46+Edge79+

Edge (Legacy)18IENone

Firefox for AndroidNoneiOS Safari15.4+Chrome for Android59+Android WebView59+Samsung Internet7.0+Opera Mobile43+

The Handle Fetch algorithm is the entry point for the fetch handling handed to the service worker context.

Input: request, a request; controller, a fetch controller; useHighResPerformanceTimers, a boolean
Output: response, a response

Let handleFetchFailed be false.
Let respondWithEntered be false.
Let eventCanceled be false.
Let response be null.
Let registration be null.
Let client be request’s client.
Let reservedClient be request’s reserved client.
Let preloadResponse be a new promise.
Let workerRealm be null.
Let eventHandled be null.
Let timingInfo be a new service worker timing info.
Let fetchInstance be the instance of the fetch algorithm representing the ongoing fetch.
Assert: request’s destination is not "serviceworker".
If request’s destination is either "embed" or "object", then:
1. Return null.
Else if request is a non-subresource request, then:

Note: If the non-subresource request is under the scope of a service worker registration, application cache is completely bypassed regardless of whether the non-subresource request uses the service worker registration.
1. If reservedClient is not null and is an environment settings object, then:
  1. If reservedClient is not a secure context, return null.
2. Else:
  1. If request’s url is not a potentially trustworthy URL, return null.
3. If request is a navigation request and the navigation triggering it was initiated with a shift+reload or equivalent, return null.
4. Assert reservedClient is not null.
5. Let storage key be the result of running obtain a storage key given reservedClient.
6. Set registration to the result of running Match Service Worker Registration given storage key and request’s url.
7. If registration is null or registration’s active worker is null, return null.
8. If request’s destination is not "report", set reservedClient’s active service worker to registration’s active worker.
9. If request is a navigation request, registration’s navigation preload enabled flag is set, request’s method is `GET`, and registration’s active worker's set of event types to handle contains fetch, then:
  
  Note: If the above is true except registration’s active worker's set of event types to handle does not contain fetch, then the user agent may wish to show a console warning, as the developer’s intent isn’t clear.
  1. Let preloadRequest be the result of cloning the request request.
  2. Let preloadRequestHeaders be preloadRequest’s header list.
  3. Let preloadResponseObject be a new Response object associated with a new Headers object whose guard is "immutable".
  4. Append to preloadRequestHeaders a new header whose name is `Service-Worker-Navigation-Preload` and value is registration’s navigation preload header value.
  5. Set preloadRequest’s service-workers mode to "none".
  6. Let preloadFetchController be null.
  7. Run the following substeps in parallel, but abort when controller’s state is "terminated" or "aborted":
    1. Set preloadFetchController to the result of fetching preloadRequest.
      
      To process response for navigationPreloadResponse, run these substeps:
      1. If navigationPreloadResponse’s type is "error", reject preloadResponse with a TypeError and terminate these substeps.
      2. Associate preloadResponseObject with navigationPreloadResponse.
      3. Resolve preloadResponse with preloadResponseObject.
  8. If aborted, then abort preloadFetchController.
10. Else, resolve preloadResponse with undefined.
Note: From this point, the service worker client starts to use its active service worker’s containing service worker registration.
Else if request is a subresource request, then:
1. If client’s active service worker is non-null, set registration to client’s active service worker’s containing service worker registration.
2. Else, return null.
Let activeWorker be registration’s active worker.
Let shouldSoftUpdate be true if any of the following are true, and false otherwise:
- request is a non-subresource request.
- request is a subresource request and registration is stale.
If the result of running the Should Skip Event algorithm with "fetch" and activeWorker is true, then:
1. If shouldSoftUpdate is true, then in parallel run the Soft Update algorithm with registration.
2. Return null.
If useHighResPerformanceTimers is true, then set useHighResPerformanceTimers to activeWorker’s global object's cross-origin isolated capability.
Let timingInfo’s start time be the coarsened shared current time given useHighResPerformanceTimers.
If activeWorker’s state is "activating", wait for activeWorker’s state to become "activated".
If the result of running the Run Service Worker algorithm with activeWorker is failure, then set handleFetchFailed to true.
Else:
1. Set workerRealm to the relevant realm of the activeWorker’s global object.
2. Set eventHandled to a new promise in workerRealm.
3. Queue a task task to run the following substeps:
  1. Let e be the result of creating an event with FetchEvent.
  2. Let requestObject be a new Request object associated with request and a new associated Headers object whose guard is "immutable".
  3. Initialize e’s type attribute to fetch.
  4. Initialize e’s cancelable attribute to true.
  5. Initialize e’s request attribute to requestObject.
  6. Initialize e’s preloadResponse to preloadResponse.
  7. Initialize e’s clientId attribute to client’s id.
  8. If request is a non-subresource request and request’s destination is not "report", initialize e’s resultingClientId attribute to reservedClient’s id, and to the empty string otherwise.
  9. If request is a navigation request, initialize e’s replacesClientId attribute to request’s replaces client id, and to the empty string otherwise.
  10. Initialize e’s handled to eventHandled.
  11. Let timingInfo’s fetch event dispatch time to the coarsened shared current time given useHighResPerformanceTimers.
  12. Dispatch e at activeWorker’s global object.
  13. Invoke Update Service Worker Extended Events Set with activeWorker and e.
  14. If e’s respond-with entered flag is set, set respondWithEntered to true.
  15. If e’s wait to respond flag is set, then:
    1. Wait until e’s wait to respond flag is unset.
    2. If e’s respond-with error flag is set, set handleFetchFailed to true.
    3. Else, set response to e’s potential response.
  16. If response is null, request’s body is not null, and request’s body's source is null, then:
    1. If request’s body is unusable, set handleFetchFailed to true.
    2. Else, cancel request’s body with undefined.
  17. If response is not null, then set response’s service worker timing info to timingInfo.
  18. If e’s canceled flag is set, set eventCanceled to true.
  19. If controller state is "terminated" or "aborted", then queue a task to signal abort on requestObject’s signal.
  If task is discarded, set handleFetchFailed to true.
  
  The task must use activeWorker’s event loop and the handle fetch task source.
Wait for task to have executed or for handleFetchFailed to be true.
If shouldSoftUpdate is true, then in parallel run the Soft Update algorithm with registration.
If respondWithEntered is false, then:
1. If eventCanceled is true, then:
  1. If eventHandled is not null, then reject eventHandled with a "NetworkError" DOMException in workerRealm.
  2. Return a network error.
2. If eventHandled is not null, then resolve eventHandled.
3. Return null.
If handleFetchFailed is true, then:
1. If eventHandled is not null, then reject eventHandled with a "NetworkError" DOMException in workerRealm.
2. Return a network error.
If eventHandled is not null, then resolve eventHandled.
Return response.

Should Skip Event

Input: eventName, a string; serviceWorker, a service worker
Output: a boolean

Note: To avoid unnecessary delays, this specification permits skipping event dispatch when no event listeners for the event have been deterministically added in the service worker’s global during the very first script execution.

If serviceWorker’s set of event types to handle does not contain eventName, then the user agent may return true.
Return false.

Fire Functional Event

Input: eventName, a string; eventConstructor, an event constructor that extends ExtendableEvent; registration, a service worker registration; initialization, optional property initialization for event, constructed from eventConstructor; postDispatchSteps, optional steps to run on the active worker's event loop, with dispatchedEvent set to the instance of eventConstructor that was dispatched.
Output: None

Assert: registration’s active worker is not null.
Let activeWorker be registration’s active worker.
If the result of running Should Skip Event with eventName and activeWorker is true, then:
1. If registration is stale, then in parallel run the Soft Update algorithm with registration.
2. Return.
If activeWorker’s state is "activating", wait for activeWorker’s state to become "activated".
If the result of running the Run Service Worker algorithm with activeWorker is failure, then:
1. If registration is stale, then in parallel run the Soft Update algorithm with registration.
2. Return.
Queue a task task to run these substeps:
1. Let event be the result of creating an event with eventConstructor and the relevant realm of activeWorker’s global object.
2. If initialization is not null, then initialize event with initialization.
3. Dispatch event on activeWorker’s global object.
4. Invoke Update Service Worker Extended Events Set with activeWorker and event.
5. If postDispatchSteps is not null, then run postDispatchSteps passing event as dispatchedEvent.
The task must use activeWorker’s event loop and the handle functional event task source.
Wait for task to have executed or been discarded.
If registration is stale, then in parallel run the Soft Update algorithm with registration.

To fire an "amazingthing" event (which is of type AmazingThingEvent) on a particular serviceWorkerRegistration, and initialize the event object’s properties, the prose would be:

Fire Functional Event "amazingthing" using AmazingThingEvent on serviceWorkerRegistration with the following properties:

propertyName

value

anotherPropertyName

anotherValue

Then run the following steps with dispatchedEvent:
1. Do whatever you need to with dispatchedEvent on the service worker’s event loop.

Note that the initialization steps and post-dispatch steps are optional. If they aren’t needed, the prose would be:

Fire Functional Event "whatever" using ExtendableEvent on serviceWorkerRegistration.

Handle Service Worker Client Unload

The user agent must run these steps when a service worker client unloads by unloading or terminating.

Input: client, a service worker client
Output: None

Run the following steps atomically.
Let registration be the service worker registration used by client.
If registration is null, abort these steps.
If any other service worker client is using registration, abort these steps.
If registration is unregistered, invoke Try Clear Registration with registration.
Invoke Try Activate with registration.

Handle User Agent Shutdown

Input: None
Output: None

For each (storage key, scope) → registration of registration map:
1. If registration’s installing worker installingWorker is not null, then:
  1. If registration’s waiting worker is null and registration’s active worker is null, invoke Clear Registration with registration and continue to the next iteration of the loop.
  2. Else, set installingWorker to null.
2. If registration’s waiting worker is not null, run the following substep in parallel:
  1. Invoke Activate with registration.

Update Service Worker Extended Events Set

Input: worker, a service worker; event, an event
Output: None

Assert: event’s dispatch flag is unset.
For each item of worker’s set of extended events:
1. If item is not active, remove item from worker’s set of extended events.
If event is active, append event to worker’s set of extended events.

Unregister

Input: job, a job
Output: none

If the origin of job’s scope url is not job’s client's origin, then:
1. Invoke Reject Job Promise with job and "SecurityError" DOMException.
2. Invoke Finish Job with job and abort these steps.
Let registration be the result of running Get Registration given job’s storage key and job’s scope url.
If registration is null, then:
1. Invoke Resolve Job Promise with job and false.
2. Invoke Finish Job with job and abort these steps.
Remove registration map[(registration’s storage key, job’s scope url)].
Invoke Resolve Job Promise with job and true.
Invoke Try Clear Registration with registration.

Note: If Try Clear Registration does not trigger Clear Registration here, Clear Registration is tried again when the last client using the registration is unloaded or the extend lifetime promises for the registration’s service workers settle.
Invoke Finish Job with job.

Set Registration

Input: storage key, a storage key; scope, a URL; updateViaCache, an update via cache mode
Output: registration, a service worker registration

Run the following steps atomically.
Let scopeString be serialized scope with the exclude fragment flag set.
Let registration be a new service worker registration whose storage key is set to storage key, scope url is set to scope, and update via cache mode is set to updateViaCache.
Set registration map[(storage key, scopeString)] to registration.
Return registration.

Clear Registration

Input: registration, a service worker registration
Output: None

Run the following steps atomically.
If registration’s installing worker is not null, then:
1. Terminate registration’s installing worker.
2. Run the Update Worker State algorithm passing registration’s installing worker and "redundant" as the arguments.
3. Run the Update Registration State algorithm passing registration, "installing" and null as the arguments.
If registration’s waiting worker is not null, then:
1. Terminate registration’s waiting worker.
2. Run the Update Worker State algorithm passing registration’s waiting worker and "redundant" as the arguments.
3. Run the Update Registration State algorithm passing registration, "waiting" and null as the arguments.
If registration’s active worker is not null, then:
1. Terminate registration’s active worker.
2. Run the Update Worker State algorithm passing registration’s active worker and "redundant" as the arguments.
3. Run the Update Registration State algorithm passing registration, "active" and null as the arguments.

Try Clear Registration

Input: registration, a service worker registration
Output: None

Invoke Clear Registration with registration if no service worker client is using registration and all of the following conditions are true:
- registration’s installing worker is null or the result of running Service Worker Has No Pending Events with registration’s installing worker is true.
- registration’s waiting worker is null or the result of running Service Worker Has No Pending Events with registration’s waiting worker is true.
- registration’s active worker is null or the result of running Service Worker Has No Pending Events with registration’s active worker is true.

Update Registration State

Input: registration, a service worker registration; target, a string (one of "installing", "waiting", and "active"); source, a service worker or null
Output: None

Let registrationObjects be an array containing all the ServiceWorkerRegistration objects associated with registration.
If target is "installing", then:
1. Set registration’s installing worker to source.
2. For each registrationObject in registrationObjects:
  1. Queue a task to set the installing attribute of registrationObject to null if registration’s installing worker is null, or the result of getting the service worker object that represents registration’s installing worker in registrationObject’s relevant settings object.
Else if target is "waiting", then:
1. Set registration’s waiting worker to source.
2. For each registrationObject in registrationObjects:
  1. Queue a task to set the waiting attribute of registrationObject to null if registration’s waiting worker is null, or the result of getting the service worker object that represents registration’s waiting worker in registrationObject’s relevant settings object.
Else if target is "active", then:
1. Set registration’s active worker to source.
2. For each registrationObject in registrationObjects:
  1. Queue a task to set the active attribute of registrationObject to null if registration’s active worker is null, or the result of getting the service worker object that represents registration’s active worker in registrationObject’s relevant settings object.
The task must use registrationObject’s relevant settings object's responsible event loop and the DOM manipulation task source.

Update Worker State

Input: worker, a service worker; state, a service worker state
Output: None

Assert: state is not "parsed".

Note: "parsed" is the initial state. A service worker is never updated to this state.
Set worker’s state to state.
Let settingsObjects be all environment settings objects whose origin is worker’s script url's origin.
For each settingsObject of settingsObjects, queue a task on settingsObject’s responsible event loop in the DOM manipulation task source to run the following steps:
1. Let objectMap be settingsObject’s service worker object map.
2. If objectMap[worker] does not exist, then abort these steps.
3. Let workerObj be objectMap[worker].
4. Set workerObj’s state to state.
5. Fire an event named statechange at workerObj.

Notify Controller Change

Input: client, a service worker client
Output: None

Assert: client is not null.
If client is an environment settings object, queue a task to fire an event named controllerchange at the ServiceWorkerContainer object that client is associated with.

The task must use client’s responsible event loop and the DOM manipulation task source.

Match Service Worker Registration

Input: storage key, a storage key; clientURL, a URL
Output: A service worker registration

Run the following steps atomically.
Let clientURLString be serialized clientURL.
Let matchingScopeString be the empty string.
Let scopeStringSet be an empty list.
For each (entry storage key, entry scope) → registration of registration map:
1. If storage key equals entry storage key, then append entry scope to the end of scopeStringSet.
Set matchingScopeString to the longest value in scopeStringSet which the value of clientURLString starts with, if it exists.

Note: The URL string matching in this step is prefix-based rather than path-structural. E.g. a client URL string with "https://example.com/prefix-of/resource.html" will match a registration for a scope with "https://example.com/prefix". The URL string comparison is safe for the same-origin security as HTTP(S) URLs are always serialized with a trailing slash at the end of the origin part of the URLs.
Let matchingScope be null.
If matchingScopeString is not the empty string, then:
1. Set matchingScope to the result of parsing matchingScopeString.
2. Assert: matchingScope’s origin and clientURL’s origin are same origin.
Return the result of running Get Registration given storage key and matchingScope.

Get Registration

Input: storage key, a storage key; scope, a URL
Output: A service worker registration

Run the following steps atomically.
Let scopeString be the empty string.
If scope is not null, set scopeString to serialized scope with the exclude fragment flag set.
For each (entry storage key, entry scope) → registration of registration map:
1. If storage key equals entry storage key and scopeString matches entry scope, then return registration.
Return null.

Get Newest Worker

Input: registration, a service worker registration
Output: newestWorker, a service worker

Run the following steps atomically.
Let newestWorker be null.
If registration’s installing worker is not null, set newestWorker to registration’s installing worker.
Else if registration’s waiting worker is not null, set newestWorker to registration’s waiting worker.
Else if registration’s active worker is not null, set newestWorker to registration’s active worker.
Return newestWorker.

Service Worker Has No Pending Events

Input: worker, a service worker
Output: True or false, a boolean

For each event of worker’s set of extended events:
1. If event is active, return false.
Return true.

Create Client

Input: client, a service worker client
Output: clientObject, a Client object

Let clientObject be a new Client object.
Set clientObject’s service worker client to client.
Return clientObject.

Create Window Client

Input: client, a service worker client; frameType, a string; visibilityState, a string; focusState, a boolean; ancestorOriginsList, a list
Output: windowClient, a WindowClient object

Let windowClient be a new WindowClient object.
Set windowClient’s service worker client to client.
Set windowClient’s frame type to frameType.
Set windowClient’s visibility state to visibilityState.
Set windowClient’s focus state to focusState.
Set windowClient’s ancestor origins array to a frozen array created from ancestorOriginsList.
Return windowClient.

Get Frame Type

Input: browsingContext, a browsing context
Output: frameType, a string

Return the value by switching on the type of browsingContext:

Nested browsing context

"nested"

Auxiliary browsing context

"auxiliary"

Otherwise

"top-level"

Resolve Get Client Promise

Input: client, a service worker client; promise, a promise
Output: none

If client is an environment settings object, then:
1. If client is not a secure context, queue a task to reject promise with a "SecurityError" DOMException, on promise’s relevant settings object's responsible event loop using the DOM manipulation task source, and abort these steps.
Else:
1. If client’s creation URL is not a potentially trustworthy URL, queue a task to reject promise with a "SecurityError" DOMException, on promise’s relevant settings object's responsible event loop using the DOM manipulation task source, and abort these steps.
If client is an environment settings object and is not a window client, then:
1. Let clientObject be the result of running Create Client algorithm with client as the argument.
2. Queue a task to resolve promise with clientObject, on promise’s relevant settings object's responsible event loop using the DOM manipulation task source, and abort these steps.
Else:
1. Let browsingContext be null.
2. If client is an environment settings object, set browsingContext to client’s global object's browsing context.
3. Else, set browsingContext to client’s target browsing context.
4. Queue a task to run the following steps on browsingContext’s event loop using the user interaction task source:
  1. Let frameType be the result of running Get Frame Type with browsingContext.
  2. Let visibilityState be browsingContext’s active document's visibilityState attribute value.
  3. Let focusState be the result of running the has focus steps with browsingContext’s active document as the argument.
  4. Let ancestorOriginsList be the empty list.
  5. If client is a window client, set ancestorOriginsList to browsingContext’s active document's relevant global object's Location object’s ancestor origins list's associated list.
  6. Queue a task to run the following steps on promise’s relevant settings object's responsible event loop using the DOM manipulation task source:
    1. If client’s discarded flag is set, resolve promise with undefined and abort these steps.
    2. Let windowClient be the result of running Create Window Client with client, frameType, visibilityState, focusState, and ancestorOriginsList.
    3. Resolve promise with windowClient.

Query Cache

Input: requestQuery, a request; options, a CacheQueryOptions object, optional; targetStorage, a request response list, optional
Output: resultList, a request response list

Let resultList be an empty list.
Let storage be null.
If the optional argument targetStorage is omitted, set storage to the relevant request response list.
Else, set storage to targetStorage.
For each requestResponse of storage:
1. Let cachedRequest be requestResponse’s request.
2. Let cachedResponse be requestResponse’s response.
3. If Request Matches Cached Item with requestQuery, cachedRequest, cachedResponse, and options returns true, then:
  1. Let requestCopy be a copy of cachedRequest.
  2. Let responseCopy be a copy of cachedResponse.
  3. Add requestCopy/responseCopy to resultList.
Return resultList.

Request Matches Cached Item

Input: requestQuery, a request; request, a request; response, a response or null, optional, defaulting to null; options, a CacheQueryOptions object, optional
Output: a boolean

If options["ignoreMethod"] is false and request’s method is not `GET`, return false.
Let queryURL be requestQuery’s url.
Let cachedURL be request’s url.
If options["ignoreSearch"] is true, then:
1. Set cachedURL’s query to the empty string.
2. Set queryURL’s query to the empty string.
If queryURL does not equal cachedURL with the exclude fragment flag set, then return false.
If response is null, options["ignoreVary"] is true, or response’s header list does not contain `Vary`, then return true.
Let fieldValues be the list containing the elements corresponding to the field-values of the Vary header for the value of the header with name `Vary`.
For each fieldValue in fieldValues:
1. If fieldValue matches "*", or the combined value given fieldValue and request’s header list does not match the combined value given fieldValue and requestQuery’s header list, then return false.
Return true.

Batch Cache Operations

Input: operations, a list of cache batch operation objects
Output: resultList, a request response list

Let cache be the relevant request response list.
Let backupCache be a new request response list that is a copy of cache.
Let addedItems be an empty list.
Try running the following substeps atomically:
1. Let resultList be an empty list.
2. For each operation in operations:
  1. If operation’s type matches neither "delete" nor "put", throw a TypeError.
  2. If operation’s type matches "delete" and operation’s response is not null, throw a TypeError.
  3. If the result of running Query Cache with operation’s request, operation’s options, and addedItems is not empty, throw an "InvalidStateError" DOMException.
  4. Let requestResponses be an empty list.
  5. If operation’s type matches "delete", then:
    1. Set requestResponses to the result of running Query Cache with operation’s request and operation’s options.
    2. For each requestResponse in requestResponses:
      1. Remove the item whose value matches requestResponse from cache.
  6. Else if operation’s type matches "put", then:
    1. If operation’s response is null, throw a TypeError.
    2. Let r be operation’s request's associated request.
    3. If r’s url's scheme is not one of "http" and "https", throw a TypeError.
    4. If r’s method is not `GET`, throw a TypeError.
    5. If operation’s options is not null, throw a TypeError.
    6. Set requestResponses to the result of running Query Cache with operation’s request.
    7. For each requestResponse of requestResponses:
      1. Remove the item whose value matches requestResponse from cache.
    8. Append operation’s request/operation’s response to cache.
    9. If the cache write operation in the previous two steps failed due to exceeding the granted quota limit, throw a "QuotaExceededError" DOMException.
    10. Append operation’s request/operation’s response to addedItems.
  7. Append operation’s request/operation’s response to resultList.
3. Return resultList.
And then, if an exception was thrown, then:
1. Remove all the items from the relevant request response list.
2. For each requestResponse of backupCache:
  1. Append requestResponse to the relevant request response list.
3. Throw the exception.
Note: When an exception is thrown, the implementation does undo (roll back) any changes made to the cache storage during the batch operation job.

Is Async Module

Input: record, a Module Record; moduleMap, a module map; base, a URL; seen, a set of URLs
Output: a boolean

If record is not a Cyclic Module Record, then:
1. Return false.
If record.[[Async]] is true, then:
1. Return true.
For each string requested of record.[[RequestedModules]]:
1. Let url be the result of resolving a module specifier given base and requested.
2. Assert: url is never failure, because resolving a module specifier must have been previously successful with these same two arguments.
3. If seen does not contain url, then:
  1. Append url to seen.
  2. If moduleMap[url] does not have a record, then:
    1. Return false.
  3. If Is Async Module for moduleMap[url]'s record, moduleMap, base, and seen is true, then:
    1. Return true.
Return false.

Appendix B: Extended HTTP headers

Service Worker Script Request

An HTTP request to fetch a service worker's script resource will include the following header:

`Service-Worker`

Indicates this request is a service worker's script resource request.

Note: This header helps administrators log the requests and detect threats.

Service Worker Script Response

An HTTP response to a service worker's script resource request can include the following header:

`Service-Worker-Allowed`

Indicates the user agent will override the path restriction, which limits the maximum allowed scope url that the script can control, to the given value.

Note: The value is a URL. If a relative URL is given, it is parsed against the script’s URL.

Default scope:

// Maximum allowed scope defaults to the path the script sits in
// "/js/" in this example
navigator.serviceWorker.register("/js/sw.js").then(() => {
  console.log("Install succeeded with the default scope '/js/'.");
});

Upper path without Service-Worker-Allowed header:

// Set the scope to an upper path of the script location
// Response has no Service-Worker-Allowed header
navigator.serviceWorker.register("/js/sw.js", { scope: "/" }).catch(() => {
  console.error("Install failed due to the path restriction violation.");
});

Upper path with Service-Worker-Allowed header:

// Set the scope to an upper path of the script location
// Response included "Service-Worker-Allowed : /"
navigator.serviceWorker.register("/js/sw.js", { scope: "/" }).then(() => {
  console.log("Install succeeded as the max allowed scope was overriden to '/'.");
});

A path restriction voliation even with Service-Worker-Allowed header:

// Set the scope to an upper path of the script location
// Response included "Service-Worker-Allowed : /foo"
navigator.serviceWorker.register("/foo/bar/sw.js", { scope: "/" }).catch(() => {
  console.error("Install failed as the scope is still out of the overriden maximum allowed scope.");
});

Syntax

ABNF for the values of the headers used by the service worker's script resource requests and responses:

Service-Worker = %x73.63.72.69.70.74 ; "script", case-sensitive

Note: The validation of the Service-Worker-Allowed header’s values is done by URL parsing algorithm (in Update algorithm) instead of using ABNF.

8. Acknowledgements

Deep thanks go to Andrew Betts for organizing and hosting a small workshop of like-minded individuals including: Jake Archibald, Jackson Gabbard, Tobie Langel, Robin Berjon, Patrick Lauke, Christian Heilmann. From the clarity of the day’s discussions and the use-cases outlined there, much has become possible. Further thanks to Andrew for raising consciousness about the offline problem. His organization of EdgeConf and inclusion of Offline as a persistent topic there has created many opportunities and connections that have enabled this work to progress.

Anne van Kesteren has generously lent his encyclopedic knowledge of Web Platform arcana and standards development experience throughout the development of the service worker. This specification would be incomplete without his previous work in describing the real-world behavior of URLs, HTTP Fetch, Promises, and DOM. Similarly, this specification would not be possible without Ian Hickson’s rigorous Web Worker spec. Much thanks to him.

In no particular order, deep gratitude for design guidance and discussion goes to: Jungkee Song, Alec Flett, David Barrett-Kahn, Aaron Boodman, Michael Nordman, Tom Ashworth, Kinuko Yasuda, Darin Fisher, Jonas Sicking, Jesús Leganés Combarro, Mark Christian, Dave Hermann, Yehuda Katz, François Remy, Ilya Grigorik, Will Chan, Domenic Denicola, Nikhil Marathe, Yves Lafon, Adam Barth, Greg Simon, Devdatta Akhawe, Dominic Cooney, Jeffrey Yasskin, Joshua Bell, Boris Zbarsky, Matt Falkenhagen, Tobie Langel, Gavin Peters, Ben Kelly, Hiroki Nakagawa, Jake Archibald, Josh Soref, Jinho Bang, Yutaka Hirano, Michael(tm) Smith, isonmad, Ali Alabbas, Philip Jägenstedt, Mike Pennisi, and Eric Willigers.

Jason Weber, Chris Wilson, Paul Kinlan, Ehsan Akhgari, and Daniel Austin have provided valuable, well-timed feedback on requirements and the standardization process.

The authors would also like to thank Dimitri Glazkov for his scripts and formatting tools which have been essential in the production of this specification. The authors are also grateful for his considerable guidance.

Thanks also to Vivian Cromwell, Greg Simon, Alex Komoroske, Wonsuk Lee, and Seojin Kim for their considerable professional support.

Service Workers

Abstract

Status of this document

1. Motivations

2. Model

2.1. Service Worker

2.1.1. Lifetime

2.1.2. Events

2.2. Service Worker Timing

2.3. Service Worker Registration

2.3.1. Lifetime

2.4. Service Worker Client

2.5. Control and Use

2.5.1. The window client case

2.5.2. The worker client case

2.6. Task Sources

2.7. User Agent Shutdown

3. Client Context

3.1. ServiceWorker

3.1.1. Getting ServiceWorker instances

3.1.2. scriptURL

3.1.3. state

3.1.4. postMessage(message, transfer)

3.1.5. postMessage(message, options)

3.1.6. Event handler

3.2. ServiceWorkerRegistration

3.2.1. Getting ServiceWorkerRegistration instances

3.2.2. installing

3.2.3. waiting

3.2.4. active

3.2.5. navigationPreload

3.2.6. scope

3.2.7. updateViaCache

3.2.8. update()

3.2.9. unregister()

3.2.10. Event handler

3.3. navigator.serviceWorker

3.4. ServiceWorkerContainer

3.4.1. controller

3.4.2. ready

3.4.3. register(scriptURL, options)

3.4.4. getRegistration(clientURL)

3.4.5. getRegistrations()

3.4.6. startMessages()

3.4.7. Event handlers

3.5. Events

3.6. NavigationPreloadManager

3.6.1. enable()

3.6.2. disable()

3.6.3. setHeaderValue(value)

3.6.4. getState()

4. Execution Context

4.1. ServiceWorkerGlobalScope

4.1.1. clients

4.1.2. registration

4.1.3. serviceWorker

4.1.4. skipWaiting()

4.1.5. Event handlers

4.2. Client

4.2.1. url

4.2.2. frameType

4.2.3. id

4.2.4. type

4.2.5. postMessage(message, transfer)

4.2.6. postMessage(message, options)

4.2.7. visibilityState

4.2.8. focused

4.2.9. ancestorOrigins

4.2.10. focus()

4.2.11. navigate(url)

4.3. Clients

4.3.1. get(id)

4.3.2. matchAll(options)

4.3.3. openWindow(url)

4.3.4. claim()

4.4. ExtendableEvent

4.4.1. event.waitUntil(f)

4.5. FetchEvent

4.5.1. event.request

4.5.2. event.preloadResponse

3.1. `ServiceWorker`

3.1.1. Getting `ServiceWorker` instances

3.1.2. `scriptURL`

3.1.3. `state`

3.1.4. `postMessage(message, transfer)`

3.1.5. `postMessage(message, options)`

3.2. `ServiceWorkerRegistration`

3.2.1. Getting `ServiceWorkerRegistration` instances

3.2.2. `installing`

3.2.3. `waiting`

3.2.4. `active`

3.2.5. `navigationPreload`

3.2.6. `scope`

3.2.7. `updateViaCache`

3.2.8. `update()`

3.2.9. `unregister()`

3.3. `navigator.serviceWorker`

3.4. `ServiceWorkerContainer`

3.4.1. `controller`

3.4.2. `ready`

3.4.3. `register(scriptURL, options)`

3.4.4. `getRegistration(clientURL)`

3.4.5. `getRegistrations()`

3.4.6. `startMessages()`

3.6. `NavigationPreloadManager`

3.6.1. `enable()`

3.6.2. `disable()`

3.6.3. `setHeaderValue(value)`

3.6.4. `getState()`

4.1. `ServiceWorkerGlobalScope`

4.1.1. `clients`

4.1.2. `registration`

4.1.3. `serviceWorker`

4.1.4. `skipWaiting()`

4.2. `Client`

4.2.1. `url`

4.2.2. `frameType`

4.2.3. `id`

4.2.4. `type`

4.2.5. `postMessage(message, transfer)`

4.2.6. `postMessage(message, options)`

4.2.7. `visibilityState`

4.2.8. `focused`

4.2.9. `ancestorOrigins`

4.2.10. `focus()`

4.2.11. `navigate(url)`

4.3. `Clients`

4.3.1. `get(id)`

4.3.2. `matchAll(options)`

4.3.3. `openWindow(url)`

4.3.4. `claim()`

4.4. `ExtendableEvent`

4.4.1. `event.waitUntil(f)`

4.5. `FetchEvent`

4.5.1. `event.request`

4.5.2. `event.preloadResponse`

4.5.3. `event.clientId`

4.5.4. `event.resultingClientId`

4.5.5. `event.replacesClientId`

4.5.6. `event.handled`

4.5.7. `event.respondWith(r)`

4.6. `ExtendableMessageEvent`

4.6.1. `event.data`

4.6.2. `event.origin`

4.6.3. `event.lastEventId`

4.6.4. `event.source`

4.6.5. `event.ports`

5.3. `self.caches`

5.3.1. `caches`

5.4. `Cache`

5.4.1. `match(request, options)`

5.4.2. `matchAll(request, options)`

5.4.3. `add(request)`

5.4.4. `addAll(requests)`

5.4.5. `put(request, response)`

5.4.6. `delete(request, options)`

5.4.7. `keys(request, options)`

5.5. `CacheStorage`

5.5.1. `match(request, options)`

5.5.2. `has(cacheName)`