Proxy fetch requests through the Background Sync API
Made with ❤ at @outlandish
Fetch Sync allows you to proxy fetch requests through the Background Sync API so that they are honoured if made when the UA is offline! Hooray!
Check out a live demo here.
- Make requests offline that will be sent when the UA regains connectivity (even if the web page is no longer open).
- Responses are forwarded back to the client as soon as they are received.
- Implements a familiar fetch-like API: similar function signature and the same return type (a Response).
- Make named requests that have their response stored in an IDBStore which you can collect in subsequent user sessions.
- Manage sync operations with
fetchSync.{get,getAll,cancel,cancelAll}()
. - Can be used with existing Service Worker infrastructures with
importScripts
, or handles SW registration for you. - If the browser does not support Background Sync, the library will fall back on normal
fetch
requests.
npm install fetch-sync --save
- Requirements
- Support
- Initialisation
- Usage
- Sync API
- Dependencies
- Test
- Development
- Contributing
- Author & License
The library utilises some new technologies so currently only works in some browsers. It definitely works in
Chrome Canary
with the experimental-web-platform-features
flag enabled.
The browser must support:
- Background Sync
- Service Worker
- Fetch
- IndexedDB
- [Promise] (https://developer.mozilla.org/en/docs/Web/JavaScript/Reference/Global_Objects/Promise)
Chrome Canary | Chrome | Firefox | IE | Opera | Safari |
---|---|---|---|---|---|
✔ | ✔ | ✘ | ✘ | ✘ | ✘ |
Existing Service Worker
If your application already uses a Service Worker, you can import the fetch-sync worker using importScripts
:
importScripts('node_modules/fetch-sync/dist/fetch-sync.sw.min.js')
And then call fetchSync.init()
somewhere in your application's initialisation procedure.
No Service Worker
fetch-sync can handle registration if you don't use a SW already...
Either serve the fetch-sync worker file with a header "Service-Worker-Allowed : /"
, or to avoid configuring headers,
create a Service Worker script in the root of your project and use the method above for 'Existing Service Worker'.
Then see the example under Usage for the fetchSync.init()
method.
Initialise fetchSync.
- options {Object} (optional) options object
Look at the documentation for sw-register
available options and for more details on Service Worker registration.
Example:
// Import client lib...
// ES6
import fetchSync from 'fetch-sync'
// ES5
var fetchSync = require('fetch-sync')
<!-- Script, using bundled dist -->
<script src="/node_modules/fetch-sync/dist/fetch-sync.min.js"></script>
// Initialise, passing in worker lib location...
fetchSync.init({
url: 'node_modules/fetch-sync/dist/fetch-sync.sw.js',
scope: '<website address>' // e.g. 'http://localhost:8000'
})
Perform a sync
Background Sync operation.
- [name] {String} (optional) name of the sync operation
- request {String|Request} URL or an instance of fetch Request
- [options] {Object} (optional) fetch options object
Returns a Promise that resolves on success of the fetch request.
If called with a name
:
- the response will be stored and can be retrieved later using e.g.
fetchSync.get('name').then(sync => sync.response)
. - the response will not automatically be removed from the IDBStore in the worker. You should request
that a named sync be removed manually by using
sync.remove()
. - see the Sync API for more details.
Examples:
-
named GET
fetchSync('GetMessages', '/messages') .then((response) => response.json()) .then((json) => console.log(json.foo))
-
unnamed POST
const post = fetchSync('/update-profile', { method: 'POST', body: { name: '' } }) // cancel the sync... post.cancel()
-
unnamed with options
const headers = new Headers(); headers.append('Authorization', 'Basic abcdefghijklmnopqrstuvwxyz'); // `fetchSync` accepts the same args as `fetch`... fetchSync('/send-message', { headers })
-
named with options
fetchSync('/get-messages', { headers })
-
unnamed with Request
fetchSync( new Request('/messages') )
Get a sync by its name.
- name {String} name of the sync operation to get
Returns a Promise that resolves with success of the sync operation.
There are also some properties/methods on the Sync. See the Sync API for more details.
Example:
fetchSync('SendMessage', '/message', { body: 'Hello, World!' })
const sync = fetchSync.get('SendMessage')
sync.then((response) => {
if (response.ok) {
alert(`Your message was sent at ${new Date(sync.syncedOn).toDateString()}.`
} else {
alert('Message failed to send.')
}
})
Get all sync operations.
Returns an array of all sync operations (named and unnamed).
Example:
fetchSync.getAll()
.then((syncs) => syncs.forEach(sync => sync.cancel()))
Cancel the sync with the given name
.
- name {String} name of the sync operation to cancel
Example:
fetchSync('Update', '/update', { body })
fetchSync.cancel('Update')
Cancel all syncs, named and unnamed.
Cancels the sync operation.
Returns a Promise of success of the cancellation.
Example:
const sync = fetchSync.get('Update')
sync.cancel()
The unique ID of the sync operation. This will be its name if it has one.
The name of the sync operation if it has one.
The time that the sync operation was created.
The time that the sync operation was completed.
Useful for named syncs that you want to retrieve later on.
As the library depends on Service Workers and no headless browser has (good enough) support for Service Workers that would allow tests to be executed within the console, tests are ran through the browser using Mocha and Chai.
On running npm test
an Express server will be started at localhost:8000
.
Run the tests:
$ cd fetch-sync
$ npm test
The library is bundled by Webpack and transpiled by Babel.
- Install dependencies:
npm install
- Start Webpack in a console:
npm run watch
- Start the test server in another:
npm test
- Navigate to
http://localhost:8000
All pull requests and issues welcome!
If you're not sure how, check out Kent C. Dodds' great video tutorials on egghead.io!
fetch-sync
was created by Sam Gluck and is released under the MIT license.