Building a Chrome Extension that adds a devtools panel. That part is working. The dev tools panel needs to be able to grab information from the current page.
The only way I see how to do it, is using a content script. Ideally I don't want to require the "tabs" or "" permissions. I'm fine with them being optional permissions".
My content scripts that I inject only appear to work on the pages that I have specified in my manifest file. I want them to work on any page the user inspects so my panel can display information.
User clicks a button in panel - I request optional permissions for "tabs" and "" in my panel.js
my devtools page .js checks for "tabs" and "" permission.
If permission is granted, I send background.js a message - load this content script.
Recieving the message doesn't seem to matter on Background.js's end as the content script loads automatically regardless on URL's specified in the manifest. It does not however appear on pages you inspect.
CODE
(in order of how it was explained above)
Button in panel.html
<button id="load">Load Dev Info</button>
https://github.com/williamspiro/HubSpot-Developer-Extension/blob/c23b8cad6f1787264bb3e8b3673c594a7db5fe7e/panel.html#L8
Permission request in Panel.js
document.querySelector('#load').addEventListener('click', function(event) {
// Permissions must be requested from inside a user gesture, like a button's
// click handler.
chrome.permissions.request({
permissions: ['tabs'],
origins:['<all_urls>']
}, function(granted) {
// The callback argument will be true if the user granted the permissions.
if (granted) {
console.log("Perm granted");
} else {
console.log("Perm denied");
}
});
});
https://github.com/williamspiro/HubSpot-Developer-Extension/blob/c23b8cad6f1787264bb3e8b3673c594a7db5fe7e/panel.js#L2-L18
permissions check if granted - sendMessage with content script name to background JS and devtools.js(loaded on devtools.html)
name: "devtools-page"
});
backgroundPageConnection.onMessage.addListener(function (message) {
// Handle responses from the background page, if any
});
console.log("devtools.js ran");
chrome.permissions.contains({
permissions: ['tabs'],
origins: ['<all_urls>']
}, function(result) {
if (result) {
// The extension has the permissions.
// Relay the tab ID to the background page
chrome.runtime.sendMessage({
tabId: chrome.devtools.inspectedWindow.tabId,
scriptToInject: "hsInspector.js"
});
} else {
// The extension doesn't have the permissions.
}
});
https://github.com/williamspiro/HubSpot-Developer-Extension/blob/c23b8cad6f1787264bb3e8b3673c594a7db5fe7e/devtools.js#L7-L35
Receive message to background.js - inject content script
chrome.runtime.onConnect.addListener(function(devToolsConnection) {
// assign the listener function to a variable so we can remove it later
var devToolsListener = function(message, sender, sendResponse) {
// Inject a content script into the identified tab
console.log("script:",message.scriptToInject);
chrome.tabs.executeScript(message.tabId,
{ file: message.scriptToInject });
}
// add the listener
devToolsConnection.onMessage.addListener(devToolsListener);
devToolsConnection.onDisconnect.addListener(function() {
devToolsConnection.onMessage.removeListener(devToolsListener);
});
});
https://github.com/williamspiro/HubSpot-Developer-Extension/blob/c23b8cad6f1787264bb3e8b3673c594a7db5fe7e/background.js#L38-L52
It appears the background JS part does not matter, because in manifest I have the content script mentioned in the manifest causing it to get injected on all URL's that match but not inspected pages.
Permissions and content scripts in manifest.json
"permissions": [
"storage",
"activeTab"
],
"optional_permissions": [
"https://app.hubspot.com/*",
"https://app.hubspotqa.com/*",
"tabs",
"<all_urls>"
],
"web_accessible_resources": [
"mac-text-cursor.svg",
"mac-text-cursor.png"
],
"content_scripts": [{
"run_at": "document_idle",
"js": ["jquery-3.2.1.min.js", "design-manager.js","hsInspector.js"],
"css": ["hsDarkIde.css"],
"document_idle": "document_start",
"matches": ["https://app.hubspot.com/*", "https://app.hubspotqa.com/*"]
}],
"content_security_policy": "script-src 'self' https://ssl.google-analytics.com; object-src 'self'",
https://github.com/williamspiro/HubSpot-Developer-Extension/blob/66ee912d56b88be052182a91257af28a4e356558/manifest.json#L25-L46
The content script I am trying to inject from the devtools is hsInspector.js
I only care to have this file get loaded if the devtools injects it. It doesn't need to load on the "matches" pages at all.
EXPECTED AND ACTUAL RESULTS
Expected result is that when inspecting a page, if you go to the extension's dev tools tab, and have the permissions required, the content script will be injected into the page.
[if that were working my goal would be to then make the content script do this]
Content script will be used to grab data from the page, message background.js the required data, background.js messages panel.js the data, panel js would render what it needs to in the panel.
Actual result: content script is not loading into the inspected page, but is instead loading onto the "matches" URLs specified under content_scripts.
If you have questions please ask. I am not an expert at building chrome extensions, so maybe I'm making this way more complicated than it needs to be. Really appreciate any help. Thank you in advance.
Related
I've taken a look at other related SO posts and the solutions haven't helped solve my issue. This is my first chrome extension, so please bear with me!
I'm writing a simple chrome extension that searches for user provided keywords on a webpage. I can't get the content script that returns the DOM content to run. Some of the code, I've taken from an answer in another SO post, but I can't seem to get it to work for me.
I put a console.log("hello world") at the top of the file, and it doesn't show up, so I think it might be the structure of my project.
manifest.json
{
"name": "keyword search",
"version": "0.0.1",
"manifest_version": 2,
"permissions": [ "tabs" , "storage", "activeTab", "<all_urls>"],
"browser_action": {
"default_popup": "html/form.html"
},
"content_scripts": [{
"matches": [ "<all_urls>" ],
"js": [ "js/jquery.min.js", "content_scripts/content_script.js" ]
}],
"homepage_url": "http://google.com/"
}
js/popup.js
function run() {
running = true;
console.log('running');
var url = "https://www.stackoverflow.com/"
// Get KW & category for search
chrome.storage.local.get(["kw"],
function (data) {
kw = data.kw;
console.log("redirecting to find kw: " + kw);
// Send current tab to url
chrome.tabs.query({active: true, currentWindow: true}, function(tabs) {
chrome.tabs.update(tabs[0].id, {url: url});
chrome.tabs.sendMessage(tabs[0].id, {type: 'DOM_request'}, searchDOM);
});
}
);
}
function searchDOM(domContent) {
console.log("beginning dom search \n" + domContent);
}
content_scripts/content_script.js
// Listen for messages
console.log("hello world")
chrome.runtime.onMessageExternal.addListener(function (msg, sender, sendResponse) {
// If the received message has the expected format...
if (msg.type === 'DOM_request') {
// Call the specified callback, passing
// the web-page's DOM content as argument
sendResponse(document.all[0].outerHTML);
}
});
console
running
redirecting to find kw: TestKeyword
beginning dom search
undefined
First, onMessageExternal is the wrong event (it's for external messaging):
you should use the standard onMessage.
Second, chrome extensions API is asynchronous so it only registers a job, returns immediately to continue to the next statement in your code without waiting for the job to complete:
chrome.tabs.update enqueues a navigation to a new URL
chrome.tabs.sendMessage enqueues a message sending job
the current page context in the tab gets destroyed along with the running content scripts
the tab starts loading the new URL
the message is delivered into the tab but there are no listeners,
but this step may instead run right after step 2 depending on various factors so the content script running in the old page will receive it which is not what you want
the tab loads the served HTML and emits a DOMContentLoaded event
your content scripts run shortly after that because of the default "run_at": "document_idle"
There are at least three methods to properly time it all:
make your content script emit a message and add an onMessage listener in the popup
use chrome.tabs.onUpdated to wait for the tab to load
use chrome.tabs.onUpdated + chrome.tabs.executeScript to simplify the entire thing
Let's take the executeScript approach.
remove "content_scripts" from manifest.json
instead of chrome.tabs.query (it's not needed) use the following:
chrome.tabs.update({url}, tab => {
chrome.tabs.onUpdated.addListener(function onUpdated(tabId, change, updatedTab) {
if (tabId === tab.id && change.status === 'complete') {
chrome.tabs.onUpdated.removeListener(onUpdated);
chrome.tabs.executeScript(tab.id, {
code: 'document.documentElement.innerHTML',
}, results => {
searchDOM(results[0]);
});
}
});
});
I currently have indexedDB on google.com domain. i want to be able to read it from google chrome extension. how can i accomplish this? do i need to add any specific permissions?
i currently have:
"permissions": [ "tabs", "bookmarks", "unlimitedStorage", "*://*/*", "identity", "https://*.google.com/*", "https://ssl.gstatic.com/", "https://www.googleapis.com/", "https://accounts.google.com/" ],
with what command i can do this? thank you!
Edit: i have readed i can access it from content script(aslong as the tab with domain is open - which is my case), but i dont know how to do that...
To access indexeddb of current tab add "activeTab" to "permissions" tag in manifest.json, Then create a content script, content script will be helpful in accessing the indexeddb as it runs in context of webpages, then add the content script created to the "content_scripts" tag in manifest.json file.
For Eg in manifest.json add the following:
"permissions": ["activeTab"],
"content_scripts": [
{
"matches": ["add the domains of the webpages where content script needs to run"],
"js": ["contentScript.js"]
}
]
For more info on matches check out here: https://developer.chrome.com/extensions/match_patterns
.
Inside content script add open the store and then perform transaction on the object store and perform queries on the object store.
For Eg in content script add following:
if (!('indexedDB' in window)) {
alert("This browser doesn't support IndexedDB");
} else {
let indexdb = window.indexedDB.open('firebaseLocalStorageDb', 1);
indexdb.onsuccess = function () {
let db = indexdb.result;
let transaction = db.transaction('firebaseLocalStorage', 'readwrite');
let storage = transaction.objectStore('firebaseLocalStorage');
console.log(storage.getAll());
};
}
Explanation of the above code:
It accesses the window object and opens the store "firebaseLocalStorageDb" with version "1", then after successfully accessing the object it looks for the result and performs transaction on the objectstore "firebaseLocalStorage" residing inside the store. Finally query the instance of objectstore "storage" to get all the key-value pairs.
For more info check: https://javascript.info/indexeddb
For anyone still interested, my solution to this problem -
this is placed in content script of extension -
chrome.extension.onConnect.addListener(function(port) {
if(port.name == "extension_request" ) {
port.onMessage.addListener(function(msg) {
if (msg.db) {
window.indexedDB.webkitGetDatabaseNames().onsuccess = function(sender,args)
{
var r = sender.target.result;
if(r.contains(msg.db)){
var openRequest = indexedDB.open(msg.db);
// your code
port.postMessage({foo: bar}); // your result which you want to send
}
}
}
}
}
and this is for background or popup script -
chrome.tabs.query({active: true, currentWindow: true}, function(tabs) {
var port = chrome.tabs.connect(tabs[0].id,{name: "extension_request"});
port.postMessage({db: "database_name_example"}); // send database name
port.onMessage.addListener(function(msg) {
if (msg.foo ) {
// do your stuff in extension
}
}
}
UPDATE
From what I can tell, it is impossible to send a message from the background script to the content script using the "sendMessage" function. However there is a horrible workaround,
In your content script's window.onload, send a message to the background script:
chrome.runtime.sendMessage( { action: "messaging", window: "app" }, this.listenForFutureMessages );
Also in the content script, have the following function:
listenForFutureMessages: function(someAction)
{
//Take some action based on the message
//If we want the background script to be able to contact
//us again, we need to give them another callback. This
//is because Chrome only allows one use per callback
chrome.runtime.sendMessage( { action: "messaging", window: "app" }, this.listenForFutureMessages );
},
In the background script, have a listener that does something like this:
chrome.runtime.onMessage.addListener(
function(request, sender, sendResponse)
{
if ( request.action === "messaging" )
{
//Save the callback for later
this.listeners[ request.window ] = sendResponse;
//Tell chrome we will be using the callback later
return true;
}
}
);
When your background script wants to send the content script a message, simply call it like this:
this.listeners[ "app" ]( { someProperty: "some value" } );
This is a stupid way to do this, but it makes this actually possible. Hope this helps anyone else who needs this functionality.
ORIGINAL
I'm unable to send a message from my background script to a content script. When I try to find the tab id, it tells me I don't have permissions even though my app has that permission. And when I receive a message from the content script, and print out the sender object, it shows tab.id = -1. The API to send a message to a content script requires a tab id!
chrome.tabs.sendMessage(integer tabId, any message, function responseCallback)
The error:
chrome.tabs is not available: You do not have permission to access this API. Ensure that the required permission or manifest property is included in your manifest.json.
Error in event handler for 'undefined': Cannot call method 'sendMessage' of undefined TypeError: Cannot call method 'sendMessage' of undefined
at chrome-extension://panoaieakcofaegcjfbmhndaekfgpijh/scripts/background.js:109:16
at Event.dispatchToListener (event_bindings:356:21)
at Event.dispatch_ (event_bindings:342:27)
at Event.dispatch (event_bindings:362:17)
at miscellaneous_bindings:167:33
at Event.dispatchToListener (event_bindings:356:21)
at Event.dispatch_ (event_bindings:342:27)
at Event.dispatch (event_bindings:362:17)
at Object.chromeHidden.Port.dispatchOnMessage (miscellaneous_bindings:253:22)
So how do I contact my content script? (I have multiple windows and need to be able to contact them individually)
My manifest:
{
"manifest_version": 2,
"name": "App",
"description": "App",
"version": "0.75",
"minimum_chrome_version": "27",
"offline_enabled": true,
"icons":
{
"16": "images/icon16.png",
"48": "images/icon48.png",
"128": "images/icon128.png"
},
"app":
{
"background":
{
"scripts":
[
"scripts/background.js"
]
}
},
"permissions":
[
"unlimitedStorage",
"fullscreen",
{
"fileSystem":
[
"write"
]
},
"background",
"<all_urls>",
"tabs"
],
"update_url": "http://192.168.1.121/app.xml"
}
There's not such a thing called "Content scripts" in a Chrome app. Your manifest file looks like a mixture of a Chrome extension. Open chrome://extensions/, enable developer mode, and you would see a warning that the "background" and "tabs" permissions are invalid for a Chrome app.
If you're implementing a Chrome app, just use chrome.runtime.sendMessage and chrome.runtime.onMessage. These messages can be send from and to your event page and the main page. For example:
// event page (aka background page)
chrome.app.runtime.onLaunched.addListener(function() {
chrome.app.window.create('main.html');
});
// Later, when you want to notify the app window
chrome.runtime.sendMessage(" ... any message ... ");
<!-- main.html -->
<script src="main.js"></script>
// main.js
chrome.runtime.onMessage.addListener(function(message, sender, sendResponse) {
// Do something with the message
});
I've been trying to communicate from my Content Script to a background page (for XHR), but then, I've been failing to even establish a communication.
Here's a snippet from content_script.js
$('.genericStreamStory').each(function(){
var link = $(this).find('.uiStreamSource a').attr('href');
$(this).find('.uiStreamFooter').append("<span class='a1' style='color:red !important;'> ยท Sample</span>");
document.querySelector('.a1').onclick=function(){
//alert('span clicked');
chrome.extension.sendMessage({method: "getHTML", data: 'hello'});
};//end of anonymous function
});
back.js
chrome.extension.onMessage.addListener(function(request, sender, sendResponse) {
if (request.method === "getHTML") {
console.log('check.. '+request.data);
}
});
and finally, manifest.json
{
"name": "App1",
"version": "0.1",
"description": "Yay, I'm so useless.",
"content_scripts": [
{
"matches": ["https://*.facebook.com/*"],
"js": [
"/js/external/jquery.js",
"/js/content_script.js"
]
}
],
"background": {
"scripts": ["/js/back.js"]
},
"manifest_version": 2
}
So this extension basically tries to create append a 'Span' after every facebook post.
When I click the Span element, I get the message in the alert box (which I've commented now). Ideally, this action should communicate with my back.js page.
In back.js, I'm logging the message sent by the content script to the console.
But then, I'm literally getting nothing in the console screen!
I've tried the following :
I replaced onMessage & sendMessage with onRequest & sendRequest. Got a PORT Error
Also tried loading an external script from a .js file replacing inline script. - Blank console!
Is there any bug in my code? Where am I making a mistake?
I am trying to show a simple desktop notification code from a content script, but it doesn't seem to work.. I have added the permissions in the maifest.json file. Is there a restriction on showing them from the content script ?
You can't show notifications directly through a content script.
But, you can show them through the background page.
Your manifest.js should look something like this:
{
"name": "Notify This",
"version": "0.1",
"permissions": [
"notifications"
],
"background_page": "background.html",
"content_scripts": [
{
"matches": ["http://www.example.com/*"],
"js": ["contentscript.js"]
}
]
}
Then use the chrome.extension.sendRequest():
// in your contentscript.js
chrome.extension.sendRequest({msg: "Sup?"}, function(response) { // optional callback - gets response
console.log(response.returnMsg);
});
And on the receiving end you should have a onRequest listener:
// in your background.html
chrome.extension.onRequest.addListener(
function(request, sender, sendResponse) {
// Create a simple text notification:
var notify = webkitNotifications.createNotification(
'48.png', // icon url - can be relative
'Hello!', // notification title
request.msg // notification body text
);
notify.show();
setTimeout(function(){ notify.cancel(); },5000);
sendResponse({returnMsg: "All good!"}); // optional response
});
Yes, notifications use Chrome specific API, and the content script is only valid for general javascript etc... The background page is where all chrome specific API's are capable of running... First you'll need to register your background page in the manifest.json file - like this:
"background_page": "background.html",
Also in the manifest file, Allow the required permissions:
"permissions": [ "notifications" ],
Then your script in the background page should look like this :
<script>
setTimeout("setNotification();",1);
function setNotification(){
var n
if (window.webkitNotifications.checkPermission() != 0){
setNotification();
return false;
}
n = window.webkitNotifications.createHTMLNotification('http://www.your-notification-address.com');
n.show();}
</script>