

const URL = "https://8014-35-223-70-178.ngrok-free.app/"; // 1
const taskChannel = new BroadcastChannel('task-channel'); // 2
taskChannel.onmessage = occasion => { // 3
persistTask(occasion.information.information); // 4
registration.sync.register('task-sync'); // 5
};
let db = null; // 6
let request = indexedDB.open("TaskDB", 1); // 7
request.onupgradeneeded = perform(occasion) { // 8
db = occasion.goal.end result; // 9
if (!db.objectStoreNames.comprises("duties")) { // 10
let tasksObjectStore = db.createObjectStore("duties", { autoIncrement: true }); // 11
}
};
request.onsuccess = perform(occasion) { db = occasion.goal.end result; }; // 12
request.onerror = perform(occasion) { console.log("Error in db: " + occasion); }; // 13
persistTask = perform(process){ // 14
let transaction = db.transaction("duties", "readwrite");
let tasksObjectStore = transaction.objectStore("duties");
let addRequest = tasksObjectStore.add(process);
addRequest.onsuccess = perform(occasion){ console.log("Process added to DB"); };
addRequest.onerror = perform(occasion) { console.log(“Error: “ + occasion); };
}
self.addEventListener('sync', async perform(occasion) { // 15
if (occasion.tag == 'task-sync') {
occasion.waitUntil(new Promise((res, rej) => { // 16
let transaction = db.transaction("duties", "readwrite");
let tasksObjectStore = transaction.objectStore("duties");
let cursorRequest = tasksObjectStore.openCursor();
cursorRequest.onsuccess = perform(occasion) { // 17
let cursor = occasion.goal.end result;
if (cursor) {
let process = cursor.worth; // 18
fetch(URL + 'todos/add', // a
{ technique: 'POST',
headers: { 'Content material-Sort': 'software/json' },
physique: JSON.stringify({ "process" : process })
}).then((serverResponse) => {
console.log("Process saved to backend.");
deleteTasks(); // b
res(); // b
}).catch((err) => {
console.log("ERROR: " + err);
rej(); //c
})
}
}
}))
}
})
async perform deleteTasks() { // 19
const transaction = db.transaction("duties", "readwrite");
const tasksObjectStore = transaction.objectStore("duties");
tasksObjectStore.clear();
await transaction.full;
}
Now let’s discuss what is going on on this code.
- We have to route our requests by way of the identical safe tunnel we created with
ngrok, so we save the URL right here. - Create the printed channel with the identical title so we will pay attention for messages.
- Right here, we’re expecting
task-channelmessage occasions. In responding to those occasions, we do two issues: - Name
persistTask()to avoid wasting the brand new process toIndexedDB. - Register a brand new
syncoccasion. That is what invokes the particular functionality for retrying requests intelligently. The sync handler permits us to specify a promise that it’s going to retry when the community is on the market, and implements a again off technique and give-up situations. - With that performed, we create a reference for our database object.
- Acquire a “request” for the deal with on our database. Every thing on
IndexedDBis dealt with asynchronously. (For a superb overview ofIndexedDB, I like to recommend this collection.) - The
onupgradeneededoccasion fires if we’re accessing a brand new or up-versioned database. - Inside
onupgradeneeded, we get a deal with on the database itself, with our internationaldbobject. - If the duties assortment shouldn’t be current, we create the duties assortment.
- If the database was efficiently created, we put it aside to our
dbobject. - Log the error if the database creation failed.
- The
persistTask()perform referred to as by the add-task broadcast occasion (4). This merely places the brand new process worth within the duties assortment. - Our sync occasion. That is referred to as by the printed occasion (5). We examine for the
occasion.tagdiscipline beingtask-syncso we all know it’s our task-syncing occasion. occasion.waitUntil()permits us to inform theserviceWorkerthat we’re not performed till thePromiseinside it completes. As a result of we’re in a sync occasion, this has particular that means. Specifically, if ourPromisefails, the syncing algorithm will hold making an attempt. Additionally, keep in mind that if the community is unavailable, it would wait till it turns into obtainable.- We outline a brand new
Promise, and inside it we start by opening a connection to the database.
- We outline a brand new
- Inside the database
onsuccesscallback, we get hold of a cursor and use it to seize the duty we saved. (We’re leveraging our wrappingPromiseto cope with nested asynchronous calls.) - Now we now have a variable with the worth of our broadcast process in it. With that in hand:
- We difficulty a brand new
fetchrequest to ourexpressJS /todos/addendpoint. - Discover that if the request succeeds, we delete the duty from the database and name
res()to resolve our outer promise. - If the request fails, we name
rej(). This may reject the containing promise, letting the Sync API know the request have to be retried.
- We difficulty a brand new
- The
deleteTasks()helper technique deletes all of the duties within the database. (This can be a simplified instance that assumes onedutiescreation at a time.)
Clearly, there’s a lot to this, however the reward is having the ability to effortlessly retry requests within the background every time our community is spotty. Keep in mind, we’re getting this within the browser, throughout every kind of units, cell and in any other case.
Testing the PWA instance
If you happen to run the PWA now and create a to-do, it’ll be despatched to the again finish and saved. The fascinating take a look at is to open devtools (F12) and disable the community. You will discover the “Offline” possibility within the “throttling” menu of the community tab like so:


