Skip to main content
Dublin Library

The Publishing Project

Permissions policy and API

 

The Permissions API gives use the ability to request permission from the user to use a given web API.

This is an update of the 2020 posts Feature Policies and Working with Feature Policies in client-side Javascript.

Feature policies have been renamed to Permissions API and the API may have changed since the original articles were posted

In this post we'll look at the Permissions API, how it works, how it's different from other APIs like geolocation that provide their own permissions, how is this different than using Content Security Policies, and how to incorporate it into your projects.

What is the Permissions API #

The permissions policy allows you to request user permission to use a given Javascript API in your application.

Not all APIs are gated behing the permissions API. The list of available APIs and their corresponding policiesn are shown in the following table

API Policy
Background Synchronization API (should always be granted) background-sync
Clipboard API clipboard-read, clipboard-write
Geolocation API geolocation
Local Font Access API local-fonts
Media Capture and Streams API microphone, camera
Notifications API notifications
Payment Handler API payment-handler
Push API push
Sensor APIs accelerometer,
gyroscope,
magnetometer,
ambient-light-sensor
Storage Access API storage-access
Storage API persistent-storage
Web Audio Output Devices API speaker-selection
Web MIDI API midi

In the first example we use a button element in the HTML document to trigger permission request. All permissions for the APIs that use the Permissions API require user gesture, as I was researching this post, I discovered that you can't trigger the permission request programatically.

async function requestPermissions() {
  const response = await navigator.permissions.request({name: "notifications"});
  const currentPermissions = response;

  console.log(`Current permissions:`, currentPermissions);
}

const activateBtn = document.getElementById("activate");

activateBtn.addEventListener("click", requestPermissions);

We can also query the permissions state whenever we want to do something about with the API.

The navigator.permissions.query method takes an array of one or more permissions and checks the status.

navigator.permissions.query({
  name: "notifications"
}).then((result) => {
  if (result.state === "granted") {
    console.log("Permission granted")
  } else if (result.state === "prompt") {
    ("Browser will prompt for permission")
  } else {
    console.log('permission was not granted')
  }
});

Finally, we can query the status of one or more permissions. This example, taken from A Tour of the JavaScript Permissions API shows how can query and display the status of the different permissions available through the Permissions API.

I chose to use Promise.allSettled() instead of Promise.all() because I want to make sure that all promises resolve rather than have the promise return when one of the component promises rejects.

The permissionsNames constant holds the names of the permissions we want to check. In this case I've added all the permissions from the table in the earlier section of this post.

const permissionsNames = [
  "geolocation",
  "notifications",
  "push",
  "midi",
  "camera",
  "microphone",
  "background-fetch",
  "background-sync",
  "persistent-storage",
  "ambient-light-sensor",
  "accelerometer",
  "gyroscope",
  "magnetometer",
  "display-capture",
  "clipboard-read",
  "clipboard-write",
  "local-fonts",
  "payment-handler",
  "storage-access",
  "speaker-selection",
]

Inside the getAllPermissions function we create an empty arra for the permissions. We'll populate this array later in the function.

The first wrapper is Promise.allSettled(). Inside the allSettled block we create a map from the permissionsName array we've defined earlier.

Then we use a switch statement to special case the push permission to account for Chrome's idiosyncracies.

The default switch block will query each of the permissions to get its current status and push the result to the allPermissions array.

The final step is to return the allPermissions array.

We then call getAllPermissions to trigger the query process.

async function getAllPermissions() {
  const allPermissions = []
  await Promise.allSettled(
    permissionsNames.map(async (permissionName) => {
      try {
        let permission
        switch (permissionName) {
          case 'push':
            permission = await navigator.permissions.query({name: permissionName, userVisibleOnly: true})
            break
          default:
            permission = await navigator.permissions.query({name: permissionName})
        }
        console.log(permission)
        allPermissions.push({
          permissionName, state:
          permission.state
        })
      }
      catch(e){
        allPermissions.push({
          permissionName,
          state: 'error',
          errorMessage: e.toString()
        })
      }
    })
  )
  return allPermissions
}

getAllPermissions()

Older APIs and how they grant permission #

You may see older APIs like geolocation and notifications use different ways to request permissions. The old ways will still work, but they also provide a more uniform way to request permissions.

This example uses the geolocation API to request the user's location.

The old way to request permission is to use geolocation.getCurrentPosition. Whenever we use this code, the browser will ask for permission.

navigator.geolocation.getCurrentPosition(function(position) {
    console.log('Geolocation permissions granted');
    console.log('Latitude:' + position.coords.latitude);
    console.log('Longitude:' + position.coords.longitude);
});

Using the permissions API the code will query the geolocation permission and then print the status of the permission.

We also set up an onchange event to detect changes of the permission and we log any changes to the permission.

navigator.permissions
  .query({ name: "geolocation" })
  .then(function (permissionStatus) {
    console.log("geolocation permission state is ", permissionStatus.state);

    permissionStatus.onchange = function () {
      console.log("geolocation permission state has changed to ", this.state);
    };
  });

Relationship with Content Security Policy (CSP) #

CSP and the Permissions API address different aspects web security.

CSP addresses the source of content for your website while the permissions API asks the user if the app can use a given web API.

So there's no real comparison as they perform different functions.

Relationship with the Permissions-Policy #

In addition to the client-side Permissions API there is also a Permissions Policy HTTP header system that allows you to set usage policies on the server before querying them on the client.

The list of available permission you can set via the HTTP headers is available in the Directives section of MDN's Permissions-Policy page.

In that page you'll see that many of the permissions available as HTTP headers are related to Chromium's web capabilities project (also known as Project Fugu). While these APIs will only work in Chromium browsers it is interesting that they get grouped with APIs that work cross browsers.

Iframe allow syntax #

iframe elements can have their own permissions that further restrict what developers can and cannot do with iframes.

You can configure iframes using the allow attribute to indicate one or more permissions that controls what the iframe can do.

These configuration are done on a per-iframe basis so you have to configure each iframe on its own.

The following example, taken from YouTube, shows how the allow attribute works.

<iframe
  width="560" height="315"
  src="https://www.youtube.com/embed/rZJgedd2Fk4"
  title="YouTube video player"
  frameborder="0"
  allow="accelerometer;
    autoplay;
    clipboard-write;
    encrypted-media;
    gyroscope;
    picture-in-picture;
    web-share" allowfullscreen>
</iframe>

The iframe allow attribute controls whether you can use an API and is closer in intent to the HTTP headers approach.

Used in combination with the client-side API it gives developers a powerful way to ask for permission before using powerful APIs.

Edit on Github