After the recent Chrome update my extension started to fire "Unchecked runtime.lastError: Tabs cannot be edited right now (user may be dragging a tab)" when I attempt to use chrome.tabs API.
There is not much info on this issue yet, but I believe this is a browser bug. In the meantime my extension causes the chrome tabs to switch noticeably slower, that it used to be. It now takes a couple of seconds to change the tab. So I'm looking for a workaround.
Any ideas how to fix this?
The only solution that I have found so far is to put my handlers in a timeout like this:
chrome.tabs.onActivated.addListener((activeInfo) => {
setTimeout(() => {
// The old listener handler moves here
}, 100);
});
But there must be a better way, right?
You will still get errors but at least it will work
chrome.tabs.onActivated.addListener(function(activeInfo) {getActivatedTab();});
function getActivatedTab(){
chrome.tabs.query({'active': true, 'lastFocusedWindow': true}, function (tabs) {
try{
if(tabs[0]!=undefined){
console.log(tabs[0].url);
}
}
catch(err){
setTimeout(function() {
getActivatedTab();
},100);
}
})
}
create separate function outside of chrome.tabs.onActivated.addListener.
this way work for me.
function insertScript(tab) {
chrome.tabs.get(tab.tabId, function (info) {
console.log(info);
});
}
chrome.tabs.onActivated.addListener(function (tab) {
setTimeout(function () {
insertScript(tab);
}, 500);
});
This is a bug in Chrome91, as pointed by #wOxxOm.
This patch is suggested if you are making a lot of calls to e.g. xchrome.tabs.query
const ChromeWrapper = {
chromeTabsQuery: function (params, callback) {
chrome.tabs.query(params, tabs => {
if (chrome.runtime.lastError) {
setTimeout(function () {
//console.warn("Patch for xchrome.tabs.query (Chrome 91).");
ChromeWrapper.chromeTabsQuery(params, callback)
}, 100); // arbitrary delay
} else {
callback(tabs)
}
})
}
}
And then, just replace the instances of
xchrome.tabs.query(...
with:
ChromeWrapper.chromeTabsQuery(...
...Alright,
I don't have an really better solution other than setTimeout, as I don't really see where are the message/event posted after the tab done, so what I do is also add a setTimeout to do some "trial and error".
But how I do it is to override the original chrome.tabs.get so that we can have the same experience as we use it as expected way. (And easily to delete this snippet when they finally fix it some day)
Here is my code, cheers
chrome.tabs.get = function () {
const orig_get = chrome.tabs.get.bind(chrome.tabs);
return async function (tabId) {
return new Promise(
resolve => {
var tryGet = () => {
orig_get(tabId)
.then(resolve)
.catch(() => {
// console.log("retry get");
setTimeout(() => {
tryGet(tabId);
}, 50);
})
};
tryGet();
}
)
}
}();
I'm still new to javascript, I have this javascript problem from CS50 that is supposed to open a mailbox and clicking on an email is supposed to open the email. I think my on click part of the problem is right, but when I open my page and click on an email it doesnt call the open_mail() function.
I've solved that the problem is that the load_mailbox function for being asynchronous is beign called after the DOM finishes to load, so technically theres no div with the class email-box when the DOM finishes to load, but i don't know how to solve this problem, can someone help please.
document.addEventListener('DOMContentLoaded', function() {
// Use buttons to toggle between views
document.querySelector('#inbox').addEventListener('click', () => load_mailbox('inbox'));
document.querySelector('#sent').addEventListener('click', () => load_mailbox('sent'));
document.querySelector('#archived').addEventListener('click', () => load_mailbox('archive'));
document.querySelector('#compose').addEventListener('click', compose_email);
document.querySelector('#compose-form').addEventListener('submit', send_mail);
document.querySelectorAll('.email-box').forEach(function(box) {
box.addEventListener('click', function (){
open_mail();
})
});
// By default, load the inbox
load_mailbox('inbox');
});
function load_mailbox(mailbox) {
fetch(`/emails/${mailbox}`)
.then(response => response.json())
.then(emails => {
document.querySelector('#email-content').innerHTML = "";
emails.forEach(inbox_mail);
})
};
function inbox_mail(email) {
const element = document.createElement('div');
if (document.querySelector(`#email-${email.id}`) === null) {
element.id = (`email-${email.id}`);
element.className = ("email-box");
element.innerHTML = `<p>From ${email.sender}</p><p>${email.subject}</p><p>At ${email.timestamp}
</p>`;
document.querySelector('#email-content').append(element);
}
}
I´d say the easiest solution would be to put the addEventListener to a point after the elements with class .email-box are created, e.g in your .then function after inbox_mail ran for each email
.then(emails => {
document.querySelector('#email-content').innerHTML = "";
emails.forEach(inbox_mail);
document.querySelectorAll('.email-box').forEach(function(box) {
box.addEventListener('click', function (){
open_mail();
});
});
});
DOMContentLoaded will trigger when the DOM from the initial request/response was loaded. What you are doing in your fetch callback is called "DOM-Manipulation" as you create elements and append them to the DOM that has already been loaded.
I create a new window in my Chrome extension like this:
chrome.windows.create({
url: 'https://example.com',
focused: false,
state: "minimized"
}, function(hiddenWindow) {
var code = "console.log('Some JS code goes here');"
chrome.tabs.onUpdated.addListener(function(tabId, info) {
if (info.status === 'complete') {
chrome.tabs.executeScript(hiddenWindow.tabs[0].id, {
code: code
},
function(results) {
console.log(results);
});
}
});
});
It is possible somehow to do this one:
First time we create a window with one tab inside (like in my code above)
Then each time we check if this window is not closed by user
If the window still exists, then open a new tab in this window and close the previous tab.
If this window no longer exists, then do it all over again starting from #1
Will be grateful for any help and ideas!
Store the id of the window in a variable (global, for example) and use chrome.windows.get - when the window is closed the API will return an error in chrome.runtime.lastError. Also, instead of closing the previous tab it seems simpler to navigate the existing tab to a new URL.
Now, the above scheme would require us to use an elaborate cascade of callbacks at worst or Promises at best, but since it's 2019 let's use the modern async/await syntax instead with the help of Mozilla WebExtension polyfill.
let wndId;
const wndOptions = {
focused: false,
state: 'minimized',
};
const code = `(${() => {
console.log('Some JS code goes here');
}})()`;
async function openMinimized(url) {
const w =
wndId &&
await browser.windows.get(wndId, {populate: true}).catch(() => {}) ||
await browser.windows.create({url, ...wndOptions});
wndId = w.id;
const [wTab] = w.tabs;
if (wTab.url !== url) browser.tabs.update(wTab.id, {url});
return new Promise(resolve => {
browser.tabs.onUpdated.addListener(async function onUpdated(tabId, info) {
if (tabId === wTab.id && info.status === 'complete') {
browser.tabs.onUpdated.removeListener(onUpdated);
resolve(browser.tabs.executeScript(tabId, {code}));
}
});
});
}
Usage:
openMinimized('https://example.com').then(results => {
console.log(results);
});
Notes:
onUpdated listener checks tabId to process only this tab
onUpdated listener unregisters itself
code can be written as normal JS with syntax highlight inside an IIFE in a template string
How to use the polyfill: download browser-polyfill.min.js from its official repo on unpkg, save in your extension directory, and load it just like any other script in your extension, for example, as a background script in manifest.json:
"background": {
"scripts": ["browser-polyfill.min.js", "background.js"]
}
How can I check if a URL has changed in JavaScript? For example, websites like GitHub, which use AJAX, will append page information after a # symbol to create a unique URL without reloading the page. What is the best way to detect if this URL changes?
Is the onload event called again?
Is there an event handler for the URL?
Or must the URL be checked every second to detect a change?
I wanted to be able to add locationchange event listeners. After the modification below, we'll be able to do it, like this
window.addEventListener('locationchange', function () {
console.log('location changed!');
});
In contrast, window.addEventListener('hashchange',() => {}) would only fire if the part after a hashtag in a url changes, and window.addEventListener('popstate',() => {}) doesn't always work.
This modification, similar to Christian's answer, modifies the history object to add some functionality.
By default, before these modifications, there's a popstate event, but there are no events for pushstate, and replacestate.
This modifies these three functions so that all fire a custom locationchange event for you to use, and also pushstate and replacestate events if you want to use those.
These are the modifications:
(() => {
let oldPushState = history.pushState;
history.pushState = function pushState() {
let ret = oldPushState.apply(this, arguments);
window.dispatchEvent(new Event('pushstate'));
window.dispatchEvent(new Event('locationchange'));
return ret;
};
let oldReplaceState = history.replaceState;
history.replaceState = function replaceState() {
let ret = oldReplaceState.apply(this, arguments);
window.dispatchEvent(new Event('replacestate'));
window.dispatchEvent(new Event('locationchange'));
return ret;
};
window.addEventListener('popstate', () => {
window.dispatchEvent(new Event('locationchange'));
});
})();
Note, we're creating a closure, to save the old function as part of the new one, so that it gets called whenever the new one is called.
In modern browsers (IE8+, FF3.6+, Chrome), you can just listen to the hashchange event on window.
In some old browsers, you need a timer that continually checks location.hash. If you're using jQuery, there is a plugin that does exactly that.
Example
Below I undo any URL change, to keep just the scrolling:
<script type="text/javascript">
if (window.history) {
var myOldUrl = window.location.href;
window.addEventListener('hashchange', function(){
window.history.pushState({}, null, myOldUrl);
});
}
</script>
Note that above used history-API is available in Chrome, Safari, Firefox 4+, and Internet Explorer 10pp4+
window.onhashchange = function() {
//code
}
window.onpopstate = function() {
//code
}
or
window.addEventListener('hashchange', function() {
//code
});
window.addEventListener('popstate', function() {
//code
});
with jQuery
$(window).bind('hashchange', function() {
//code
});
$(window).bind('popstate', function() {
//code
});
EDIT after a bit of researching:
It somehow seems that I have been fooled by the documentation present on Mozilla docs. The popstate event (and its callback function onpopstate) are not triggered whenever the pushState() or replaceState() are called in code. Therefore the original answer does not apply in all cases.
However there is a way to circumvent this by monkey-patching the functions according to #alpha123:
var pushState = history.pushState;
history.pushState = function () {
pushState.apply(history, arguments);
fireEvents('pushState', arguments); // Some event-handling function
};
Original answer
Given that the title of this question is "How to detect URL change" the answer, when you want to know when the full path changes (and not just the hash anchor), is that you can listen for the popstate event:
window.onpopstate = function(event) {
console.log("location: " + document.location + ", state: " + JSON.stringify(event.state));
};
Reference for popstate in Mozilla Docs
Currently (Jan 2017) there is support for popstate from 92% of browsers worldwide.
With jquery (and a plug-in) you can do
$(window).bind('hashchange', function() {
/* things */
});
http://benalman.com/projects/jquery-hashchange-plugin/
Otherwise yes, you would have to use setInterval and check for a change in the hash event (window.location.hash)
Update! A simple draft
function hashHandler(){
this.oldHash = window.location.hash;
this.Check;
var that = this;
var detect = function(){
if(that.oldHash!=window.location.hash){
alert("HASH CHANGED - new has" + window.location.hash);
that.oldHash = window.location.hash;
}
};
this.Check = setInterval(function(){ detect() }, 100);
}
var hashDetection = new hashHandler();
Add a hash change event listener!
window.addEventListener('hashchange', function(e){console.log('hash changed')});
Or, to listen to all URL changes:
window.addEventListener('popstate', function(e){console.log('url changed')});
This is better than something like the code below because only one thing can exist in window.onhashchange and you'll possibly be overwriting someone else's code.
// Bad code example
window.onhashchange = function() {
// Code that overwrites whatever was previously in window.onhashchange
}
this solution worked for me:
function checkURLchange(){
if(window.location.href != oldURL){
alert("url changed!");
oldURL = window.location.href;
}
}
var oldURL = window.location.href;
setInterval(checkURLchange, 1000);
None of these seem to work when a link is clicked that which redirects you to a different page on the same domain. Hence, I made my own solution:
let pathname = location.pathname;
window.addEventListener("click", function() {
if (location.pathname != pathname) {
pathname = location.pathname;
// code
}
});
Edit: You can also check for the popstate event (if a user goes back a page)
window.addEventListener("popstate", function() {
// code
});
Best wishes,
Calculus
If none of the window events are working for you (as they aren't in my case), you can also use a MutationObserver that looks at the root element (non-recursively).
// capture the location at page load
let currentLocation = document.location.href;
const observer = new MutationObserver((mutationList) => {
if (currentLocation !== document.location.href) {
// location changed!
currentLocation = document.location.href;
// (do your event logic here)
}
});
observer.observe(
document.getElementById('root'),
{
childList: true,
// important for performance
subtree: false
});
This may not always be feasible, but typically, if the URL changes, the root element's contents change as well.
I have not profiled, but theoretically this has less overhead than a timer because the Observer pattern is typically implemented so that it just loops through the subscriptions when a change occurs. We only added one subscription here. The timer on the other hand would have to check very frequently in order to ensure that the event was triggered immediately after URL change.
Also, this has a good chance of being more reliable than a timer since it eliminates timing issues.
Although an old question, the Location-bar project is very useful.
var LocationBar = require("location-bar");
var locationBar = new LocationBar();
// listen to all changes to the location bar
locationBar.onChange(function (path) {
console.log("the current url is", path);
});
// listen to a specific change to location bar
// e.g. Backbone builds on top of this method to implement
// it's simple parametrized Backbone.Router
locationBar.route(/some\-regex/, function () {
// only called when the current url matches the regex
});
locationBar.start({
pushState: true
});
// update the address bar and add a new entry in browsers history
locationBar.update("/some/url?param=123");
// update the address bar but don't add the entry in history
locationBar.update("/some/url", {replace: true});
// update the address bar and call the `change` callback
locationBar.update("/some/url", {trigger: true});
To listen to url changes, see below:
window.onpopstate = function(event) {
console.log("location: " + document.location + ", state: " + JSON.stringify(event.state));
};
Use this style if you intend to stop/remove listener after some certain condition.
window.addEventListener('popstate', function(e) {
console.log('url changed')
});
The answer below comes from here(with old javascript syntax(no arrow function, support IE 10+)):
https://stackoverflow.com/a/52809105/9168962
(function() {
if (typeof window.CustomEvent === "function") return false; // If not IE
function CustomEvent(event, params) {
params = params || {bubbles: false, cancelable: false, detail: null};
var evt = document.createEvent("CustomEvent");
evt.initCustomEvent(event, params.bubbles, params.cancelable, params.detail);
return evt;
}
window.CustomEvent = CustomEvent;
})();
(function() {
history.pushState = function (f) {
return function pushState() {
var ret = f.apply(this, arguments);
window.dispatchEvent(new CustomEvent("pushState"));
window.dispatchEvent(new CustomEvent("locationchange"));
return ret;
};
}(history.pushState);
history.replaceState = function (f) {
return function replaceState() {
var ret = f.apply(this, arguments);
window.dispatchEvent(new CustomEvent("replaceState"));
window.dispatchEvent(new CustomEvent("locationchange"));
return ret;
};
}(history.replaceState);
window.addEventListener("popstate", function() {
window.dispatchEvent(new CustomEvent("locationchange"));
});
})();
While doing a little chrome extension, I faced the same problem with an additionnal problem : Sometimes, the page change but not the URL.
For instance, just go to the Facebook Homepage, and click on the 'Home' button. You will reload the page but the URL won't change (one-page app style).
99% of the time, we are developping websites so we can get those events from Frameworks like Angular, React, Vue etc..
BUT, in my case of a Chrome extension (in Vanilla JS), I had to listen to an event that will trigger for each "page change", which can generally be caught by URL changed, but sometimes it doesn't.
My homemade solution was the following :
listen(window.history.length);
var oldLength = -1;
function listen(currentLength) {
if (currentLength != oldLength) {
// Do your stuff here
}
oldLength = window.history.length;
setTimeout(function () {
listen(window.history.length);
}, 1000);
}
So basically the leoneckert solution, applied to window history, which will change when a page changes in a single page app.
Not rocket science, but cleanest solution I found, considering we are only checking an integer equality here, and not bigger objects or the whole DOM.
Found a working answer in a separate thread:
There's no one event that will always work, and monkey patching the pushState event is pretty hit or miss for most major SPAs.
So smart polling is what's worked best for me. You can add as many event types as you like, but these seem to be doing a really good job for me.
Written for TS, but easily modifiable:
const locationChangeEventType = "MY_APP-location-change";
// called on creation and every url change
export function observeUrlChanges(cb: (loc: Location) => any) {
assertLocationChangeObserver();
window.addEventListener(locationChangeEventType, () => cb(window.location));
cb(window.location);
}
function assertLocationChangeObserver() {
const state = window as any as { MY_APP_locationWatchSetup: any };
if (state.MY_APP_locationWatchSetup) { return; }
state.MY_APP_locationWatchSetup = true;
let lastHref = location.href;
["popstate", "click", "keydown", "keyup", "touchstart", "touchend"].forEach((eventType) => {
window.addEventListener(eventType, () => {
requestAnimationFrame(() => {
const currentHref = location.href;
if (currentHref !== lastHref) {
lastHref = currentHref;
window.dispatchEvent(new Event(locationChangeEventType));
}
})
})
});
}
Usage
observeUrlChanges((loc) => {
console.log(loc.href)
})
I created this event that is very similar to the hashchange event
// onurlchange-event.js v1.0.1
(() => {
const hasNativeEvent = Object.keys(window).includes('onurlchange')
if (!hasNativeEvent) {
let oldURL = location.href
setInterval(() => {
const newURL = location.href
if (oldURL === newURL) {
return
}
const urlChangeEvent = new CustomEvent('urlchange', {
detail: {
oldURL,
newURL
}
})
oldURL = newURL
dispatchEvent(urlChangeEvent)
}, 25)
addEventListener('urlchange', event => {
if (typeof(onurlchange) === 'function') {
onurlchange(event)
}
})
}
})()
Example of use:
window.onurlchange = event => {
console.log(event)
console.log(event.detail.oldURL)
console.log(event.detail.newURL)
}
addEventListener('urlchange', event => {
console.log(event)
console.log(event.detail.oldURL)
console.log(event.detail.newURL)
})
for Chrome 102+ (2022-05-24)
navigation.addEventListener("navigate", e => {
console.log(`navigate ->`,e.destination.url)
});
API references WICG/navigation-api
Look at the jQuery unload function. It handles all the things.
https://api.jquery.com/unload/
The unload event is sent to the window element when the user navigates away from the page. This could mean one of many things. The user could have clicked on a link to leave the page, or typed in a new URL in the address bar. The forward and back buttons will trigger the event. Closing the browser window will cause the event to be triggered. Even a page reload will first create an unload event.
$(window).unload(
function(event) {
alert("navigating");
}
);
window.addEventListener("beforeunload", function (e) {
// do something
}, false);
You are starting a new setInterval at each call, without cancelling the previous one - probably you only meant to have a setTimeout
Enjoy!
var previousUrl = '';
var observer = new MutationObserver(function(mutations) {
if (location.href !== previousUrl) {
previousUrl = location.href;
console.log(`URL changed to ${location.href}`);
}
});
Another simple way you can do this is by adding a click event, through a class name to the anchor tags on the page to detect when it has been clicked, then you can now use the window.location.href to get the url data which you can use to run your ajax request to the server. Simple and Easy.