We Keep Coding

upgrading chrome extension from manifest version2 to v3, need to get clipboard text in background.js

Hi I am converting Google chrome extension from manifest version-2 to version-3
facing 2 issues those are mentioned below, but before that I will explain what extension in expected to do.
On click specific button on webpage I am calling console application that is copying JSON string in clipboard, then in chrome extension background.js I am getting clipboard data and passing it to content.js which is showing it in web page.
Errors / challenges:
1- Need to get clipboard text into a variable in background.js. I am able to get it in content.js but I need it to get it in background.js
2- I am getting these 2 error in background.js console, but extension is working
Unchecked runtime.lastError: Native host has exited.
Unchecked runtime.lastError: The message port closed before a response was received.
My Background.js looks like this
var port = null;
var tabId = null;
chrome.runtime.onMessage.addListener((request, sender, sendResponse) => {
tabId=sender.tab.id;
var hostName = "my.console.app";
port = chrome.runtime.connectNative(hostName);
port.onDisconnect.addListener(onDisconnected);
sendResponse({status: 'ok'});
return true;
});
function onDisconnected() {
port = null;
SendResponse();
}
//this funciton need to upgrade to manifest version 3 because in v3 `chrome.extension.getBackgroundPage` is not compatible
function SendResponse() {
bg = chrome.extension.getBackgroundPage();
bg.document.body.innerHTML = ""; // clear the background page
var helper = null;
if (helper == null) {
helper = bg.document.createElement("textarea");
helper.style.position = "absolute";
helper.style.border = "none";
document.body.appendChild(helper);
}
//Focus the textarea
helper.select();
// perform a Paste in the selected control, here the textarea
bg.document.execCommand("Paste");
// Send data back to content_script
chrome.tabs.sendMessage(tabId, { action: "MY_CUSTOM_EVENT", response: helper.value });
}
content.js
document.addEventListener("MY_CUSTOM_EVENT", function (data) {
chrome.runtime.sendMessage({ runConsoleApp: true }, response => {
});
});
async function copyToTheClipboard(textToCopy){
navigator.clipboard.readText()
.then(text => {
//console.log('Pasted content: ', text);
$('.simulateEidResponse').html(text).trigger('click');
})
.catch(err => {
console.error('Failed to read clipboard contents: ', err);
});
}

Currently there's no way to do it in the background script due to crbug.com/1404835.
In the future the workaround will be execCommand + offscreen API.
The only reliable solution that works regardless of whether the tab is focused or not is to use document.execCommand("Paste") in the content script.
// background.js
chrome.runtime.onMessage.addListener((msg, sender, sendResponse) => {
if (msg.runConsoleApp) {
chrome.runtime.connectNative('my.console.app')
.onDisconnect.addListener(() => sendResponse(true));
return true;
}
});
// content.js
document.addEventListener('MY_CUSTOM_EVENT', async e => {
await chrome.runtime.sendMessage({ runConsoleApp: true });
document.querySelector('.simulateEidResponse').focus();
document.execCommand('selectAll');
document.execCommand('paste');
});

How to add back button event in Universal Windows App without using WinjS Library?

This is my main.js
(function () {
"use strict";
//No need of WinJS
var activation = Windows.ApplicationModel.Activation;
var roaming = Windows.Storage.ApplicationData.current.roamingSettings;
// For App Start Up
Windows.UI.WebUI.WebUIApplication.addEventListener("activated", function (args) {
if (args.detail[0].kind === activation.ActivationKind.launch) {
if (roaming.values["currentUri"]) {
if (roaming.values["UserName"])
{
localStorage.setItem("UserName", roaming.values["UserName"]);
window.location.href = roaming.values["currentUri"];
}
}
}
});
// For App Suspension
Windows.UI.WebUI.WebUIApplication.addEventListener("suspending", function (args) {
roaming.values["currentUri"] = window.location.href;
roaming.values["UserName"] = localStorage.getItem("UserName");
});
// For Resuming App
Windows.UI.WebUI.WebUIApplication.addEventListener("resuming", function (args) {
var roam = Windows.Storage.ApplicationData.current.roamingSettings;
if (roam) {
if (roam.values["currentUri"]) {
localStorage.setItem("UserName", roam.values["UserName"]);
window.location.href = roam.values["currentUri"];
}
}
}, false);
// not working backpressed event
Windows.UI.WebUI.WebUIApplication.addEventListener("backpressed", function (args) {
// to do
}, false);})();
I need to add back key press event for windows phone without using winjs library?
Can anyone suggest me?
I am using ms-appx-web context in my app. I dont want to use winjs library.

I need to add back key press event for windows phone without using winjs library?
The backpressed event should be attached to Windows.Phone.UI.Input.HardwareButtons but not Windows.UI.WebUI.WebUIApplication.
If you refer to HardwareButtons.BackPressed and HardwareButtons, you will find the backpressed event is used like this:
var hardwareButtons = Windows.Phone.UI.Input.HardwareButtons;
function onBackPressed(eventArgs) { /* Your code */ }
// addEventListener syntax
hardwareButtons.addEventListener("backpressed", onBackPressed);
hardwareButtons.removeEventListener("backpressed", onBackPressed);
And since you are not making a Single Page Application. This event should be attached on every new page's JS codes.
Update: If you want to know your current device programmatically, you can use the following if-statement:
if (deviceInfo.operatingSystem.toLowerCase() == "windowsphone")
{
//do your windows phone logic
} else if (deviceInfo.operatingSystem.toLowerCase() == "windows")
{
//do your windows logic
}

I used this-
var flag = Windows.Foundation.Metadata.ApiInformation.isTypePresent("Windows.Phone.UI.Input.HardwareButtons");
if (flag) {
var hardwareButtons = Windows.Phone.UI.Input.HardwareButtons;
hardwareButtons.addEventListener("backpressed", onBackPressed);
}
It worked for me well!

Firefox add-on pageshow event fires before worker able to receive messages

I'm writing a firefox plugin and keeping track of each page's workers in an array. Apart from a bit of fancy footwork required to manage this array (as described here https://bugzilla.mozilla.org/show_bug.cgi?id=686035 and here Addon SDK - context-menu and page-mod workers) everything is working properly. One issue I'm having is that when listening to the tabs pageshow event (or the worker's own pageshow event for that matter), the callback seems to fire before the worker is actually ready. When retrieving the page's corresponding worker in the callback and using it to try to send a message to the content script, I'm receiving the error The page is currently hidden and can no longer be used until it is visible again. Normally, I'd just use a setTimeout and grit my teeth, but this isn't available for add-ons. What's a suitable workaround? The code for the main part of the add-on is below:
var { ToggleButton } = require('sdk/ui/button/toggle');
var panels = require('sdk/panel');
var tabs = require('sdk/tabs');
var self = require('sdk/self');
var pageMods = require('sdk/page-mod');
var ss = require('sdk/simple-storage');
var workers = [];
ss.storage.isPluginActive = ss.storage.isPluginActive || false;
var button = ToggleButton({
id: 'tomorrowww',
label: 'Tomorowww',
icon: {
'16': './icon-16.png',
'32': './icon-32.png',
'64': './icon-64.png'
},
onChange: handleButtonChange
});
var panel = panels.Panel({
contentURL: self.data.url('panel.html'),
contentScriptFile: self.data.url('panel-script.js'),
onHide: handlePanelHide,
width: 342,
height: 270
});
panel.port.on('panel-ready', handlePanelReady);
panel.port.on('plugin-toggled', handlePluginToggled);
panel.port.on('link-clicked', handleLinkClicked);
pageMods.PageMod({
include: ['*'],
contentScriptFile: [self.data.url('CancerDOMManager.js'), self.data.url('content-script.js')],
contentStyleFile: self.data.url('content-style.css'),
onAttach: function (worker) {
addWorker(worker);
sendActiveState(ss.storage.isPluginActive);
}
});
// move between tabs
tabs.on('activate', function () {
sendActiveState();
});
// this actually fires before the worker's pageshow event so isn't useful as the workers array will be out of sync
//tabs.on('pageshow', function () {
// sendActiveState();
//});
function addWorker (worker) {
if(workers.indexOf(worker) > -1) {
return;
}
worker.on('detach', handleWorkerDetach);
worker.on('pageshow', handleWorkerShown);
worker.on('pagehide', handleWorkerHidden);
workers.push(worker);
}
function handleWorkerDetach () {
removeWorker(this, true);
}
function handleWorkerShown () {
addWorker(this);
// back / forward page history
// trying to send the state here will trigger the page hidden error
sendActiveState();
}
function handleWorkerHidden () {
removeWorker(this);
}
function removeWorker (worker, removeEvents) {
var index = workers.indexOf(worker);
removeEvents = removeEvents || false;
if(index > -1) {
if(removeEvents) {
worker.removeListener('detach', handleWorkerDetach);
worker.removeListener('pageshow', handleWorkerShown);
worker.removeListener('pagehide', handleWorkerHidden);
}
workers.splice(index, 1);
}
}
function getWorkersForCurrentTab () {
var i;
var tabWorkers = [];
i = workers.length;
while(--i > -1) {
if(workers[i].tab.id === tabs.activeTab.id) {
tabWorkers.push(workers[i]);
}
}
return tabWorkers;
}
function handlePanelReady () {
setActive(ss.storage.isPluginActive);
}
function setActive (bool) {
ss.storage.isPluginActive = bool;
panel.port.emit('active-changed', bool);
sendActiveState();
}
function sendActiveState () {
var tabWorkers = getWorkersForCurrentTab();
var i = tabWorkers.length;
while(--i > -1) {
tabWorkers[i].port.emit('toggle-plugin', ss.storage.isPluginActive);
}
}
function handleButtonChange (state) {
if(state.checked) {
panel.show({
position: button
});
}
}
function handlePanelHide () {
button.state('window', {checked: false});
}
function handleLinkClicked (url) {
if(panel.isShowing) {
panel.hide();
}
tabs.open(url);
}
function handlePluginToggled (bool) {
if(panel.isShowing) {
panel.hide();
}
setActive(bool);
}

try using contentScriptWhen: "start" in the page-mod

I was dealing with a similar problem. I think I have it working the way I want now by putting the listener in the content script instead of the addon script. I listen for the event on the window, I then emit a message from my content script to my addon script, my addon script then sends a message back to the content script with the information needed from the addon script.
In my code, I am working on update the preferences in the content script to ensure that the tab always has the most up to date settings when they are changed, only the addon script can listen to the prefs change event.
This particular snippet will listen for when the page is navigated to from history (i.e., back or forward button), will inform the addon script, the addon script will get the most up to date preferences, and then send them back to a port listening in the content script.
Content script:
window.onpageshow = function(){
console.log("onpageshow event fired (content script)");
self.port.emit("triggerPrefChange", '');
};
Addon Script (e.g., main.js:
worker.port.on("triggerPrefChange", function() {
console.log("Received request to triggerPrefChange in the addon script");
worker.port.emit("setPrefs", prefSet.prefs);
});
Since the event is being fired from the DOM event, the page must be shown. I am not sure if listening to the pageshow event in the addon script is doing what we think.

Chrome tab.create and capture screenshot of new created tab

I have an issue with capturing a newly created tab in chrome.
I create the new tab with chrome.tabs.create and pass the tabid to my callback function which captures it.
function createtab(url) {
chrome.tabs.create({'url': url,'active':false}, function(tab) {
captureWindowTab((tab.id);
});
}
function captureWindowTab(tabid) {
chrome.tabs.update(tabid, {}, function() {
chrome.tabs.captureVisibleTab(27, {format:"png"}, function(dataUrl) {
capturecallback(dataUrl);
});
});
}
function capturecallback(dataurl) {
console.log(dataurl);
}
It works ONLY when i do it on current existing tabs. i cannot get it to work on newly created tabs. always returns undefined.
I dont understand whats the issue.

Based on the documentation it seems you need the host permission (or all hosts permission) to be able to capture it. See the docuemtation at:
https://developer.chrome.com/extensions/tabs.html#method-captureVisibleTab
Do you have the host permission or all hosts permission set?
I think you might also need the "tabs" permission, if you don't already have it.

Resolved it like this:
chrome.tabs.onUpdated.addListener(function(tabid , info) {
//console.log('loading tab'+tabid);
if(info.status == "complete") {
chrome.tabs.get(tabid,function(tab) {
chrome.topSites.get(function(sites){
tab.url = NewTab.getHostFromUrl(tab.url);
console.log(tab.url);
//console.log('loaded '+tab.url);
for (var i = 8 - 1; i >= 0; i--) {
sites[i].url = NewTab.getHostFromUrl(sites[i].url);
//console.log('checking '+sites[i].url);
if(tab.url == sites[i].url && tab.url != 'newtab') {
chrome.tabs.update(tabid, {'highlighted':true,'active':true}, function(tab){
chrome.tabs.captureVisibleTab(chrome.windows.WINDOW_ID_CURRENT, {format:"jpeg","quality":30}, function(dataUrl) {
// add data URL to array
if(dataUrl) {
console.log('its a winner!');
window.localStorage['topsite_'+encodeURI(NewTab.getHostFromUrl(tab.url))+'_thumbnail'] = dataUrl;
NewTab.getTopSites();
}
});
});
}
};
});
});
}
});

Dynamic extension context menu that depends on selected text

I am trying to create entries on the Chrome context menu based on what is selected.
I found several questions about this on Stackoverflow, and for all of them the answer is: use a content script with a "mousedown" listener that looks at the current selection and creates the Context Menu.
I implemented this, but it does not always work. Sometimes all the log messages say that the context menu was modified as I wanted, but the context menu that appears is not updated.
Based on this I suspected it was a race condition: sometimes chrome starts rendering the context menu before the code ran completely.
I tried adding a eventListener to "contextmenu" and "mouseup". The later triggers when the user selects the text with the mouse, so it changes the contextmenu much before it appears (even seconds). Even with this technique, I still see the same error happening!
This happens very often in Chrome 22.0.1229.94 (Mac), occasionally in Chromium 20.0.1132.47 (linux) and it did not happen in 2 minutes trying on Windows (Chrome 22.0.1229.94).
What is happening exactly? How can I fix that? Is there any other workaround?
Here is a simplified version of my code (not so simple because I am keeping the log messages):
manifest.json:
{
"name": "Test",
"version": "0.1",
"permissions": ["contextMenus"],
"content_scripts": [{
"matches": ["http://*/*", "https://*/*"],
"js": ["content_script.js"]
}],
"background": {
"scripts": ["background.js"]
},
"manifest_version": 2
}
content_script.js
function loadContextMenu() {
var selection = window.getSelection().toString().trim();
chrome.extension.sendMessage({request: 'loadContextMenu', selection: selection}, function (response) {
console.log('sendMessage callback');
});
}
document.addEventListener('mousedown', function(event){
if (event.button == 2) {
loadContextMenu();
}
}, true);
background.js
function SelectionType(str) {
if (str.match("^[0-9]+$"))
return "number";
else if (str.match("^[a-z]+$"))
return "lowercase string";
else
return "other";
}
chrome.extension.onMessage.addListener(function(msg, sender, sendResponse) {
console.log("msg.request = " + msg.request);
if (msg.request == "loadContextMenu") {
var type = SelectionType(msg.selection);
console.log("selection = " + msg.selection + ", type = " + type);
if (type == "number" || type == "lowercase string") {
console.log("Creating context menu with title = " + type);
chrome.contextMenus.removeAll(function() {
console.log("contextMenus.removeAll callback");
chrome.contextMenus.create(
{"title": type,
"contexts": ["selection"],
"onclick": function(info, tab) {alert(1);}},
function() {
console.log("ContextMenu.create callback! Error? " + chrome.extension.lastError);});
});
} else {
console.log("Removing context menu")
chrome.contextMenus.removeAll(function() {
console.log("contextMenus.removeAll callback");
});
}
console.log("handling message 'loadContextMenu' done.");
}
sendResponse({});
});

The contextMenus API is used to define context menu entries. It does not need to be called right before a context menu is opened. So, instead of creating the entries on the contextmenu event, use the selectionchange event to continuously update the contextmenu entry.
I will show a simple example which just displays the selected text in the context menu entry, to show that the entries are synchronized well.
Use this content script:
document.addEventListener('selectionchange', function() {
var selection = window.getSelection().toString().trim();
chrome.runtime.sendMessage({
request: 'updateContextMenu',
selection: selection
});
});
At the background, we're going to create the contextmenu entry only once. After that, we update the contextmenu item (using the ID which we get from chrome.contextMenus.create).
When the selection is empty, we remove the context menu entry if needed.
// ID to manage the context menu entry
var cmid;
var cm_clickHandler = function(clickData, tab) {
alert('Selected ' + clickData.selectionText + ' in ' + tab.url);
};
chrome.runtime.onMessage.addListener(function(msg, sender, sendResponse) {
if (msg.request === 'updateContextMenu') {
var type = msg.selection;
if (type == '') {
// Remove the context menu entry
if (cmid != null) {
chrome.contextMenus.remove(cmid);
cmid = null; // Invalidate entry now to avoid race conditions
} // else: No contextmenu ID, so nothing to remove
} else { // Add/update context menu entry
var options = {
title: type,
contexts: ['selection'],
onclick: cm_clickHandler
};
if (cmid != null) {
chrome.contextMenus.update(cmid, options);
} else {
// Create new menu, and remember the ID
cmid = chrome.contextMenus.create(options);
}
}
}
});
To keep this example simple, I assumed that there's only one context menu entry. If you want to support more entries, create an array or hash to store the IDs.
Tips
Optimization - To reduce the number of chrome.contextMenus API calls, cache the relevant values of the parameters. Then, use a simple === comparison to check whether the contextMenu item need to be created/updated.
Debugging - All chrome.contextMenus methods are asynchronous. To debug your code, pass a callback function to the .create, .remove or .update methods.

MDN doc for menus.create(), 'title' param
You can use "%s" in the string. If you do this in a menu item, and some text is selected in the page when the menu is shown, then the selected text will be interpolated into the title.
https://developer.mozilla.org/en-US/docs/Mozilla/Add-ons/WebExtensions/API/menus/create
Thus
browser.contextMenus.create({
id: 'menu-search',
title: "Search '%s'", // selected text as %s
contexts: ['selection'], // show only if selection exist
})