I have JavaScript that is doing activity periodically. When the user is not looking at the site (i.e., the window or tab does not have focus), it'd be nice to not run.
Is there a way to do this using JavaScript?
My reference point: Gmail Chat plays a sound if the window you're using isn't active.
Since originally writing this answer, a new specification has reached recommendation status thanks to the W3C.βThe Page Visibility API (on MDN) now allows us to more accurately detect when a page is hidden to the user.
document.addEventListener("visibilitychange", onchange);
Current browser support:
Chrome 13+
Internet Explorer 10+
Firefox 10+
Opera 12.10+ [read notes]
The following code falls back to the less reliable blur/focus method in incompatible browsers:
(function() {
var hidden = "hidden";
// Standards:
if (hidden in document)
document.addEventListener("visibilitychange", onchange);
else if ((hidden = "mozHidden") in document)
document.addEventListener("mozvisibilitychange", onchange);
else if ((hidden = "webkitHidden") in document)
document.addEventListener("webkitvisibilitychange", onchange);
else if ((hidden = "msHidden") in document)
document.addEventListener("msvisibilitychange", onchange);
// IE 9 and lower:
else if ("onfocusin" in document)
document.onfocusin = document.onfocusout = onchange;
// All others:
else
window.onpageshow = window.onpagehide
= window.onfocus = window.onblur = onchange;
function onchange (evt) {
var v = "visible", h = "hidden",
evtMap = {
focus:v, focusin:v, pageshow:v, blur:h, focusout:h, pagehide:h
};
evt = evt || window.event;
if (evt.type in evtMap)
document.body.className = evtMap[evt.type];
else
document.body.className = this[hidden] ? "hidden" : "visible";
}
// set the initial state (but only if browser supports the Page Visibility API)
if( document[hidden] !== undefined )
onchange({type: document[hidden] ? "blur" : "focus"});
})();
onfocusin and onfocusout are required for IE 9 and lower, while all others make use of onfocus and onblur, except for iOS, which uses onpageshow and onpagehide.
I would use jQuery because then all you have to do is this:
$(window).blur(function(){
//your code here
});
$(window).focus(function(){
//your code
});
Or at least it worked for me.
There are 3 typical methods used to determine if the user can see the HTML page, however none of them work perfectly:
The W3C Page Visibility API is supposed to do this (supported since: Firefox 10, MSIE 10, Chrome 13). However, this API only raises events when the browser tab is fully overriden (e.g. when the user changes from one tab to another one). The API does not raise events when the visibility cannot be determined with 100% accuracy (e.g. Alt+Tab to switch to another application).
Using focus/blur based methods gives you a lot of false positive. For example, if the user displays a smaller window on top of the browser window, the browser window will lose the focus (onblur raised) but the user is still able to see it (so it still need to be refreshed). See also http://javascript.info/tutorial/focus
Relying on user activity (mouse move, clicks, key typed) gives you a lot of false positive too. Think about the same case as above, or a user watching a video.
In order to improve the imperfect behaviors described above, I use a combination of the 3 methods: W3C Visibility API, then focus/blur and user activity methods in order to reduce the false positive rate. This allows to manage the following events:
Changing browser tab to another one (100% accuracy, thanks to the W3C Page Visibility API)
Page potentially hidden by another window, e.g. due to Alt+Tab (probabilistic = not 100% accurate)
User attention potentially not focused on the HTML page (probabilistic = not 100% accurate)
This is how it works: when the document lose the focus, the user activity (such as mouse move) on the document is monitored in order to determine if the window is visible or not. The page visibility probability is inversely proportional to the time of the last user activity on the page: if the user makes no activity on the document for a long time, the page is most probably not visible. The code below mimics the W3C Page Visibility API: it behaves the same way but has a small false positive rate. It has the advantage to be multibrowser (tested on Firefox 5, Firefox 10, MSIE 9, MSIE 7, Safari 5, Chrome 9).
<div id="x"></div>
<script>
/**
Registers the handler to the event for the given object.
#param obj the object which will raise the event
#param evType the event type: click, keypress, mouseover, ...
#param fn the event handler function
#param isCapturing set the event mode (true = capturing event, false = bubbling event)
#return true if the event handler has been attached correctly
*/
function addEvent(obj, evType, fn, isCapturing){
if (isCapturing==null) isCapturing=false;
if (obj.addEventListener){
// Firefox
obj.addEventListener(evType, fn, isCapturing);
return true;
} else if (obj.attachEvent){
// MSIE
var r = obj.attachEvent('on'+evType, fn);
return r;
} else {
return false;
}
}
// register to the potential page visibility change
addEvent(document, "potentialvisilitychange", function(event) {
document.getElementById("x").innerHTML+="potentialVisilityChange: potentialHidden="+document.potentialHidden+", document.potentiallyHiddenSince="+document.potentiallyHiddenSince+" s<br>";
});
// register to the W3C Page Visibility API
var hidden=null;
var visibilityChange=null;
if (typeof document.mozHidden !== "undefined") {
hidden="mozHidden";
visibilityChange="mozvisibilitychange";
} else if (typeof document.msHidden !== "undefined") {
hidden="msHidden";
visibilityChange="msvisibilitychange";
} else if (typeof document.webkitHidden!=="undefined") {
hidden="webkitHidden";
visibilityChange="webkitvisibilitychange";
} else if (typeof document.hidden !=="hidden") {
hidden="hidden";
visibilityChange="visibilitychange";
}
if (hidden!=null && visibilityChange!=null) {
addEvent(document, visibilityChange, function(event) {
document.getElementById("x").innerHTML+=visibilityChange+": "+hidden+"="+document[hidden]+"<br>";
});
}
var potentialPageVisibility = {
pageVisibilityChangeThreshold:3*3600, // in seconds
init:function() {
function setAsNotHidden() {
var dispatchEventRequired=document.potentialHidden;
document.potentialHidden=false;
document.potentiallyHiddenSince=0;
if (dispatchEventRequired) dispatchPageVisibilityChangeEvent();
}
function initPotentiallyHiddenDetection() {
if (!hasFocusLocal) {
// the window does not has the focus => check for user activity in the window
lastActionDate=new Date();
if (timeoutHandler!=null) {
clearTimeout(timeoutHandler);
}
timeoutHandler = setTimeout(checkPageVisibility, potentialPageVisibility.pageVisibilityChangeThreshold*1000+100); // +100 ms to avoid rounding issues under Firefox
}
}
function dispatchPageVisibilityChangeEvent() {
unifiedVisilityChangeEventDispatchAllowed=false;
var evt = document.createEvent("Event");
evt.initEvent("potentialvisilitychange", true, true);
document.dispatchEvent(evt);
}
function checkPageVisibility() {
var potentialHiddenDuration=(hasFocusLocal || lastActionDate==null?0:Math.floor((new Date().getTime()-lastActionDate.getTime())/1000));
document.potentiallyHiddenSince=potentialHiddenDuration;
if (potentialHiddenDuration>=potentialPageVisibility.pageVisibilityChangeThreshold && !document.potentialHidden) {
// page visibility change threshold raiched => raise the even
document.potentialHidden=true;
dispatchPageVisibilityChangeEvent();
}
}
var lastActionDate=null;
var hasFocusLocal=true;
var hasMouseOver=true;
document.potentialHidden=false;
document.potentiallyHiddenSince=0;
var timeoutHandler = null;
addEvent(document, "pageshow", function(event) {
document.getElementById("x").innerHTML+="pageshow/doc:<br>";
});
addEvent(document, "pagehide", function(event) {
document.getElementById("x").innerHTML+="pagehide/doc:<br>";
});
addEvent(window, "pageshow", function(event) {
document.getElementById("x").innerHTML+="pageshow/win:<br>"; // raised when the page first shows
});
addEvent(window, "pagehide", function(event) {
document.getElementById("x").innerHTML+="pagehide/win:<br>"; // not raised
});
addEvent(document, "mousemove", function(event) {
lastActionDate=new Date();
});
addEvent(document, "mouseover", function(event) {
hasMouseOver=true;
setAsNotHidden();
});
addEvent(document, "mouseout", function(event) {
hasMouseOver=false;
initPotentiallyHiddenDetection();
});
addEvent(window, "blur", function(event) {
hasFocusLocal=false;
initPotentiallyHiddenDetection();
});
addEvent(window, "focus", function(event) {
hasFocusLocal=true;
setAsNotHidden();
});
setAsNotHidden();
}
}
potentialPageVisibility.pageVisibilityChangeThreshold=4; // 4 seconds for testing
potentialPageVisibility.init();
</script>
Since there is currently no working cross-browser solution without false positive, you should better think twice about disabling periodical activity on your web site.
Using : Page Visibility API
document.addEventListener( 'visibilitychange' , function() {
if (document.hidden) {
console.log('bye');
} else {
console.log('well back');
}
}, false );
Can i use ? http://caniuse.com/#feat=pagevisibility
I started off using the community wiki answer, but realised that it wasn't detecting alt-tab events in Chrome. This is because it uses the first available event source, and in this case it's the page visibility API, which in Chrome seems to not track alt-tabbing.
I decided to modify the script a bit to keep track of all possible events for page focus changes. Here's a function you can drop in:
function onVisibilityChange(callback) {
var visible = true;
if (!callback) {
throw new Error('no callback given');
}
function focused() {
if (!visible) {
callback(visible = true);
}
}
function unfocused() {
if (visible) {
callback(visible = false);
}
}
// Standards:
if ('hidden' in document) {
visible = !document.hidden;
document.addEventListener('visibilitychange',
function() {(document.hidden ? unfocused : focused)()});
}
if ('mozHidden' in document) {
visible = !document.mozHidden;
document.addEventListener('mozvisibilitychange',
function() {(document.mozHidden ? unfocused : focused)()});
}
if ('webkitHidden' in document) {
visible = !document.webkitHidden;
document.addEventListener('webkitvisibilitychange',
function() {(document.webkitHidden ? unfocused : focused)()});
}
if ('msHidden' in document) {
visible = !document.msHidden;
document.addEventListener('msvisibilitychange',
function() {(document.msHidden ? unfocused : focused)()});
}
// IE 9 and lower:
if ('onfocusin' in document) {
document.onfocusin = focused;
document.onfocusout = unfocused;
}
// All others:
window.onpageshow = window.onfocus = focused;
window.onpagehide = window.onblur = unfocused;
};
Use it like this:
onVisibilityChange(function(visible) {
console.log('the page is now', visible ? 'focused' : 'unfocused');
});
This version listens for all the different visibility events and fires a callback if any of them causes a change. The focused and unfocused handlers make sure that the callback isn't called multiple times if multiple APIs catch the same visibility change.
There is a neat library available on GitHub:
https://github.com/serkanyersen/ifvisible.js
Example:
// If page is visible right now
if( ifvisible.now() ){
// Display pop-up
openPopUp();
}
I've tested version 1.0.1 on all browsers I have and can confirm that it works with:
IE9, IE10
FF 26.0
Chrome 34.0
... and probably all newer versions.
Doesn't fully work with:
IE8 - always indicate that tab/window is currently active (.now() always returns true for me)
I create a Comet Chat for my app, and when I receive a message from another user I use:
if(new_message){
if(!document.hasFocus()){
audio.play();
document.title="Have new messages";
}
else{
audio.stop();
document.title="Application Name";
}
}
This is really tricky. There seems to be no solution given the following requirements.
The page includes iframes that you have no control over
You want to track visibility state change regardless of the change being triggered by a TAB change (ctrl+tab) or a window change (alt+tab)
This happens because:
The page Visibility API can reliably tell you of a tab change (even with iframes), but it can't tell you when the user changes windows.
Listening to window blur/focus events can detect alt+tabs and ctrl+tabs, as long as the iframe doesn't have focus.
Given these restrictions, it is possible to implement a solution that combines
- The page Visibility API
- window blur/focus
- document.activeElement
That is able to:
1) ctrl+tab when parent page has focus: YES
2) ctrl+tab when iframe has focus: YES
3) alt+tab when parent page has focus: YES
4) alt+tab when iframe has focus: NO <-- bummer
When the iframe has focus, your blur/focus events don't get invoked at all, and the page Visibility API won't trigger on alt+tab.
I built upon #AndyE's solution and implemented this (almost good) solution here:
https://dl.dropboxusercontent.com/u/2683925/estante-components/visibility_test1.html
(sorry, I had some trouble with JSFiddle).
This is also available on Github: https://github.com/qmagico/estante-components
This works on chrome/chromium.
It kind works on firefox, except that it doesn't load the iframe contents (any idea why?)
Anyway, to resolve the last problem (4), the only way you can do that is to listen for blur/focus events on the iframe.
If you have some control over the iframes, you can use the postMessage API to do that.
https://dl.dropboxusercontent.com/u/2683925/estante-components/visibility_test2.html
I still haven't tested this with enough browsers.
If you can find more info about where this doesn't work, please let me know in the comments below.
var visibilityChange = (function (window) {
var inView = false;
return function (fn) {
window.onfocus = window.onblur = window.onpageshow = window.onpagehide = function (e) {
if ({focus:1, pageshow:1}[e.type]) {
if (inView) return;
fn("visible");
inView = true;
} else if (inView) {
fn("hidden");
inView = false;
}
};
};
}(this));
visibilityChange(function (state) {
console.log(state);
});
http://jsfiddle.net/ARTsinn/JTxQY/
this works for me on chrome 67, firefox 67,
if(!document.hasFocus()) {
// do stuff
}
this worked for me
document.addEventListener("visibilitychange", function() {
document.title = document.hidden ? "I'm away" : "I'm here";
});
demo: https://iamsahilralkar.github.io/document-hidden-demo/
This works in all modern browsers:
when changing tabs
when changing windows(Alt+Tab)
when maximizing another program from the taskbar
var eventName;
var visible = true;
var propName = "hidden";
if (propName in document) eventName = "visibilitychange";
else if ((propName = "msHidden") in document) eventName = "msvisibilitychange";
else if ((propName = "mozHidden") in document) eventName = "mozvisibilitychange";
else if ((propName = "webkitHidden") in document) eventName = "webkitvisibilitychange";
if (eventName) document.addEventListener(eventName, handleChange);
if ("onfocusin" in document) document.onfocusin = document.onfocusout = handleChange; //IE 9
window.onpageshow = window.onpagehide = window.onfocus = window.onblur = handleChange;// Changing tab with alt+tab
// Initialize state if Page Visibility API is supported
if (document[propName] !== undefined) handleChange({ type: document[propName] ? "blur" : "focus" });
function handleChange(evt) {
evt = evt || window.event;
if (visible && (["blur", "focusout", "pagehide"].includes(evt.type) || (this && this[propName]))){
visible = false;
console.log("Out...")
}
else if (!visible && (["focus", "focusin", "pageshow"].includes(evt.type) || (this && !this[propName]))){
visible = true;
console.log("In...")
}
}
In HTML 5 you could also use:
onpageshow: Script to be run when the window becomes visible
onpagehide: Script to be run when the window is hidden
See:
https://developer.mozilla.org/en-US/docs/Web/Events/pageshow
https://developer.mozilla.org/en-US/docs/Web/Events/pagehide
u can use :
(function () {
var requiredResolution = 10; // ms
var checkInterval = 1000; // ms
var tolerance = 20; // percent
var counter = 0;
var expected = checkInterval / requiredResolution;
//console.log('expected:', expected);
window.setInterval(function () {
counter++;
}, requiredResolution);
window.setInterval(function () {
var deviation = 100 * Math.abs(1 - counter / expected);
// console.log('is:', counter, '(off by', deviation , '%)');
if (deviation > tolerance) {
console.warn('Timer resolution not sufficient!');
}
counter = 0;
}, checkInterval);
})();
A slightly more complicated way would be to use setInterval() to check mouse position and compare to last check. If the mouse hasn't moved in a set amount of time, the user is probably idle.
This has the added advantage of telling if the user is idle, instead of just checking if the window is not active.
As many people have pointed out, this is not always a good way to check whether the user or browser window is idle, as the user might not even be using the mouse or is watching a video, or similar. I am just suggesting one possible way to check for idle-ness.
This is an adaptation of the answer from Andy E.
This will do a task e.g. refresh the page every 30 seconds,
but only if the page is visible and focused.
If visibility can't be detected, then only focus will be used.
If the user focuses the page, then it will update immediately
The page won't update again until 30 seconds after any ajax call
var windowFocused = true;
var timeOut2 = null;
$(function(){
$.ajaxSetup ({
cache: false
});
$("#content").ajaxComplete(function(event,request, settings){
set_refresh_page(); // ajax call has just been made, so page doesn't need updating again for 30 seconds
});
// check visibility and focus of window, so as not to keep updating unnecessarily
(function() {
var hidden, change, vis = {
hidden: "visibilitychange",
mozHidden: "mozvisibilitychange",
webkitHidden: "webkitvisibilitychange",
msHidden: "msvisibilitychange",
oHidden: "ovisibilitychange" /* not currently supported */
};
for (hidden in vis) {
if (vis.hasOwnProperty(hidden) && hidden in document) {
change = vis[hidden];
break;
}
}
document.body.className="visible";
if (change){ // this will check the tab visibility instead of window focus
document.addEventListener(change, onchange,false);
}
if(navigator.appName == "Microsoft Internet Explorer")
window.onfocus = document.onfocusin = document.onfocusout = onchangeFocus
else
window.onfocus = window.onblur = onchangeFocus;
function onchangeFocus(evt){
evt = evt || window.event;
if (evt.type == "focus" || evt.type == "focusin"){
windowFocused=true;
}
else if (evt.type == "blur" || evt.type == "focusout"){
windowFocused=false;
}
if (evt.type == "focus"){
update_page(); // only update using window.onfocus, because document.onfocusin can trigger on every click
}
}
function onchange () {
document.body.className = this[hidden] ? "hidden" : "visible";
update_page();
}
function update_page(){
if(windowFocused&&(document.body.className=="visible")){
set_refresh_page(1000);
}
}
})();
set_refresh_page();
})
function get_date_time_string(){
var d = new Date();
var dT = [];
dT.push(d.getDate());
dT.push(d.getMonth())
dT.push(d.getFullYear());
dT.push(d.getHours());
dT.push(d.getMinutes());
dT.push(d.getSeconds());
dT.push(d.getMilliseconds());
return dT.join('_');
}
function do_refresh_page(){
// do tasks here
// e.g. some ajax call to update part of the page.
// (date time parameter will probably force the server not to cache)
// $.ajax({
// type: "POST",
// url: "someUrl.php",
// data: "t=" + get_date_time_string()+"&task=update",
// success: function(html){
// $('#content').html(html);
// }
// });
}
function set_refresh_page(interval){
interval = typeof interval !== 'undefined' ? interval : 30000; // default time = 30 seconds
if(timeOut2 != null) clearTimeout(timeOut2);
timeOut2 = setTimeout(function(){
if((document.body.className=="visible")&&windowFocused){
do_refresh_page();
}
set_refresh_page();
}, interval);
}
For a solution without jQuery check out Visibility.js which provides information about three page states
visible ... page is visible
hidden ... page is not visible
prerender ... page is being prerendered by the browser
and also convenience-wrappers for setInterval
/* Perform action every second if visible */
Visibility.every(1000, function () {
action();
});
/* Perform action every second if visible, every 60 sec if not visible */
Visibility.every(1000, 60*1000, function () {
action();
});
A fallback for older browsers (IE < 10; iOS < 7) is also available
The Chromium team is currently developing the Idle Detection API. It is available as an origin trial since Chrome 88, which is already the 2nd origin trial for this feature. An earlier origin trial went from Chrome 84 through Chrome 86.
It can also be enabled via a flag:
Enabling via chrome://flags
To experiment with the Idle Detection API locally, without an
origin trial token, enable the
#enable-experimental-web-platform-features flag in
chrome://flags.
A demo can be found here:
https://idle-detection.glitch.me/
It has to be noted though that this API is permission-based (as it should be, otherwise this could be misused to monitor a user's behaviour!).
For angular.js, here is a directive (based on the accepted answer) that will allow your controller to react to a change in visibility:
myApp.directive('reactOnWindowFocus', function($parse) {
return {
restrict: "A",
link: function(scope, element, attrs) {
var hidden = "hidden";
var currentlyVisible = true;
var functionOrExpression = $parse(attrs.reactOnWindowFocus);
// Standards:
if (hidden in document)
document.addEventListener("visibilitychange", onchange);
else if ((hidden = "mozHidden") in document)
document.addEventListener("mozvisibilitychange", onchange);
else if ((hidden = "webkitHidden") in document)
document.addEventListener("webkitvisibilitychange", onchange);
else if ((hidden = "msHidden") in document)
document.addEventListener("msvisibilitychange", onchange);
else if ("onfocusin" in document) {
// IE 9 and lower:
document.onfocusin = onshow;
document.onfocusout = onhide;
} else {
// All others:
window.onpageshow = window.onfocus = onshow;
window.onpagehide = window.onblur = onhide;
}
function onchange (evt) {
//occurs both on leaving and on returning
currentlyVisible = !currentlyVisible;
doSomethingIfAppropriate();
}
function onshow(evt) {
//for older browsers
currentlyVisible = true;
doSomethingIfAppropriate();
}
function onhide(evt) {
//for older browsers
currentlyVisible = false;
doSomethingIfAppropriate();
}
function doSomethingIfAppropriate() {
if (currentlyVisible) {
//trigger angular digest cycle in this scope
scope.$apply(function() {
functionOrExpression(scope);
});
}
}
}
};
});
You can use it like this example: <div react-on-window-focus="refresh()">, where refresh() is a scope function in the scope of whatever Controller is in scope.
Simple/immediate check:
if(document.hidden) {
// do something
}
Visibility change event:
document.addEventListener("visibilitychange", function() {
console.log(document.visibilityState); // "hidden" or "visible"
}, false);
Promise-based event:
// An `await`able function that resolves when page visibility changes:
function visibilityChange(state="") {
return new Promise(resolve => {
document.addEventListener("visibilitychange", function() {
if(!state || document.visibilityState === state) {
resolve(document.visibilityState);
document.removeEventListener("visibilitychange", arguments.callee);
}
});
});
}
// Use it like this:
await visibilityChange();
console.log(document.visibilityState);
// Or wait for page to become...
await visibilityChange("visible");
await visibilityChange("hidden");
(Note: I was the one who added the latter two solutions into this answer, since that question is now closed and I couldn't add my own answer. Just in case someone thinks I've copied them from that post without crediting.)
If you want to act on whole browser blur:
As I commented, if browser lose focus none of the suggested events fire. My idea is to count up in a loop and reset the counter if an event fire. If the counter reach a limit I do a location.href to an other page. This also fire if you work on dev-tools.
var iput=document.getElementById("hiddenInput");
,count=1
;
function check(){
count++;
if(count%2===0){
iput.focus();
}
else{
iput.blur();
}
iput.value=count;
if(count>3){
location.href="http://Nirwana.com";
}
setTimeout(function(){check()},1000);
}
iput.onblur=function(){count=1}
iput.onfocus=function(){count=1}
check();
This is a draft successful tested on FF.
I reread the #daniel-buckmaster version
I didn't make the multiple attempt, however, the code seems more elegant to me...
// on-visibility-change.js v1.0.1, based on https://stackoverflow.com/questions/1060008/is-there-a-way-to-detect-if-a-browser-window-is-not-currently-active#38710376
function onVisibilityChange(callback) {
let d = document;
let visible = true;
let prefix;
if ('hidden' in d) {
prefix = 'h';
} else if ('webkitHidden' in d) {
prefix = 'webkitH';
} else if ('mozHidden' in d) {
prefix = 'mozH';
} else if ('msHidden' in d) {
prefix = 'msH';
} else if ('onfocusin' in d) { // ie 9 and lower
d.onfocusin = focused;
d.onfocusout = unfocused;
} else { // others
window.onpageshow = window.onfocus = focused;
window.onpagehide = window.onblur = unfocused;
};
if (prefix) {
visible = !d[prefix + 'idden'];
d.addEventListener(prefix.substring(0, prefix.length - 1) + 'visibilitychange', function() {
(d[prefix + 'idden'] ? unfocused : focused)();
});
};
function focused() {
if (!visible) {
callback(visible = true);
};
};
function unfocused() {
if (visible) {
callback(visible = false);
};
};
};
Here is a solid, modern solution. (Short a sweet ππ½)
document.addEventListener("visibilitychange", () => {
console.log( document.hasFocus() )
})
This will setup a listener to trigger when any visibility event is fired which could be a focus or blur.
my code
let browser_active = ((typeof document.hasFocus != 'undefined' ? document.hasFocus() : 1) ? 1 : 0);
if (!browser_active) {
// active
}
I have emscripten compiled c++ code (using openGL) that runs in my webpage(converted to WebGL) and renders an image. Everything is working great, Except that the Emscripten compiled Javascript blocks my usage of the "delete" "backspace" "tab" keys when I type into a textarea on the webpage. Note that all the letter keys and "space" work just fine.
For reference, the Module I have instantiated in my JavaScript code is almost exactly the same as the default module:
var Module = {
preRun: [],
postRun: [],
print: (function() {
var element = document.getElementById('emscriptenOutput');
if (element) element.value = ''; // clear browser cache
return function(text) {
if (arguments.length > 1) text = Array.prototype.slice.call(arguments).join(' ');
// These replacements are necessary if you render to raw HTML
//text = text.replace(/&/g, "&");
//text = text.replace(/</g, "<");
//text = text.replace(/>/g, ">");
//text = text.replace('\n', '<br>', 'g');
console.log(text);
if (element) {
element.value += text + "\n";
element.scrollTop = element.scrollHeight; // focus on bottom
}
};
})(),
printErr: function(text) {
if (arguments.length > 1) text = Array.prototype.slice.call(arguments).join(' ');
if (0) { // XXX disabled for safety typeof dump == 'function') {
dump(text + '\n'); // fast, straight to the real console
} else {
console.error(text);
}
},
canvas: (function() {
var canvas = document.getElementById('canvas');
// As a default initial behavior, pop up an alert when webgl context is lost. To make your
// application robust, you may want to override this behavior before shipping!
// See http://www.khronos.org/registry/webgl/specs/latest/1.0/#5.15.2
canvas.addEventListener("webglcontextlost", function(e) { alert('WebGL context lost. You will need to reload the page.'); e.preventDefault(); }, false);
return canvas;
})(),
totalDependencies: 0,
//doNotCaptureKeyboard: true,
monitorRunDependencies: function(left) {
this.totalDependencies = Math.max(this.totalDependencies, left);
// Module.setStatus(left ? 'Preparing... (' + (this.totalDependencies-left) + '/' + this.totalDependencies + ')' : 'All downloads complete.');
}
};
(function() {
var memoryInitializer = 'renderer.js.mem';
if (typeof Module['locateFile'] === 'function') {
memoryInitializer = Module['locateFile'](memoryInitializer);
} else if (Module['memoryInitializerPrefixURL']) {
memoryInitializer = Module['memoryInitializerPrefixURL'] + memoryInitializer;
}
var meminitXHR = Module['memoryInitializerRequest'] = new XMLHttpRequest();
meminitXHR.open('GET', memoryInitializer, true);
meminitXHR.responseType = 'arraybuffer';
meminitXHR.send(null);
})();
var script = document.createElement('script');
script.src = "renderer.js";
document.body.appendChild(script);
Ofcourse, the emscripten compiled JavaScript is gibberish to the human eye but the culprit most likely lies in there.
My guess is that the WebGL context is eating the keyboard events, but I'm not sure. I've looked at these links and tried to include "doNotCaptureKeyboard: true" to the module element, but with no success.
https://forum.unity.com/threads/disable-enable-keyboard-in-runtime-webgl.286557/#post-1892527
https://github.com/kripken/emscripten/issues/2668#event-154218404
Anyone have any experience with the same sort of issue? I'm at a loss.
For those curious, I've got a solution but not sure if it's the best. I've just added event listeners of my own to block the event listeners of the compiled JavaScript code:
/* code to prevent emscripten compiled code from eating key input */
window.addEventListener('keydown', function(event){
event.stopImmediatePropagation();
}, true);
window.addEventListener('keyup', function(event){
event.stopImmediatePropagation();
}, true);
Inside emscripten src/library_glfw.js (around line 400, milage might vary on your version) there's a very bizarre check that goes like this:
// This logic comes directly from the sdl implementation. We cannot
// call preventDefault on all keydown events otherwise onKeyPress will
// not get called
if (event.keyCode === 8 /* backspace */ || event.keyCode === 9 /* tab */) {
event.preventDefault();
}
Now If you comment out those lines those events will get propagated correctly, I'm not sure why they've put that SDL special case inside a generic code path.
PS: This answer will probably be outdated if they ever fix that bit of
code, so if you are reading it from the future just use it as a past
reference.
I'd like to learn how to bind a CNRTL-S or COMMAND-S to call a function that I have on my page which AJAX saves the textarea content's
How can I bind those two commands? I used to use the following when it was just a textarea, but since adding TinyMCE it no longer works. Suggestions?
// Keybind the Control-Save
jQuery('#text_area_content').bind('keydown', 'ctrl+s',function (evt){
saveTextArea();
return false;
});
// Keybind the Meta-Save Mac
jQuery('#text_area_content').bind('keydown', 'meta+s',function (evt){
saveTextArea();
return false;
});
Thanks
To use a custom method for saving, i declare my saving function in the tinymce.init method
tinyMCE.init({
// General options
mode: "none",
/* some standard init params, plugins, ui, custom styles, etc */
save_onsavecallback: saveActiveEditor,
save_oncancelcallback: cancelActiveEditor
});
Then i define the function itself
function saveActiveEditor() {
var activeEditor = tinyMCE.activeEditor;
var saveUrl = "http://my.ajax.path/saveStuff";
var idEditor = activeEditor.id;
var contentEditor = activeEditor.getContent();
/* the next line is for a custom language listbox to edit different locales */
var localeEditor = activeEditor.controlManager.get('lbLanguages').selectedValue;
$.post(saveUrl ,
{ id: idEditor, content: contentEditor, locale: localeEditor },
function(results) {
if (results.Success) {
// switch back to display instead of edit
return false;
}
else {
activeEditor.windowManager.alert('Error saving data');
return false;
}
},
'json'
);
return false;
}
Don't forget to return false to override the default save action that posts back your data to the server.
edit to add: i only let the user change one tinymce instance at a time. You may want to change the locating the current instance to something else :)
edit #2: TinyMce already catches the Ctrl+s binding to process the data. Since it also cleans up html and is able to handle specific rules it's given when saving, the solution i propose is to plug your way of saving in tinyMce instead of fully overriding the Ctrl+s binding
The problem here is that the tinymce iframe does not delegate the events to the parent window. You can define custom_shortcuts in tinymce and/or use the following syntax:
// to delegate it to the parent window i use
var create_keydown_event = function(combo){
var e = { type : 'keydown' }, m = combo.split(/\+/);
for (var i=0, l=m.length; i<l; i++){
switch(m[i]){
case 'ctrl': e.metaKey = true;
case 'alt': case 'shift': e[m[i] + 'Key'] = true; break;
default : e.charCode = e.keyCode = e.which = m[i].toUpperCase().charCodeAt(0);
}
}
return e;
}
var handler = function(){
setTimeout(function(){
var e = create_keydown_event(combo);
window.parent.receiveShortCutEvent(e);
}, 1);
}
//ed.addShortcut(combo, description, handler);
ed.addShortcut('ctrl+s', 'save_shortcut', handler);
in the parent window you need a function receiveShortCutEvent which will sort out what to do with it