addEventListener polyfill for IE6+

addEventListener-polyfill.js

//addEventListener polyfill 1.0 / Eirik Backer / MIT Licence
(function(win, doc){
	if(win.addEventListener)return;		//No need to polyfill

	function docHijack(p){var old = doc[p];doc[p] = function(v){return addListen(old(v))}}
	function addEvent(on, fn, self){
		return (self = this).attachEvent('on' + on, function(e){
			var e = e || win.event;
			e.preventDefault  = e.preventDefault  || function(){e.returnValue = false}
			e.stopPropagation = e.stopPropagation || function(){e.cancelBubble = true}
			fn.call(self, e);
		});
	}
	function addListen(obj, i){
		if(i = obj.length)while(i--)obj[i].addEventListener = addEvent;
		else obj.addEventListener = addEvent;
		return obj;
	}

	addListen([doc, win]);
	if('Element' in win)win.Element.prototype.addEventListener = addEvent;			//IE8
	else{		//IE < 8
		doc.attachEvent('onreadystatechange', function(){addListen(doc.all)});		//Make sure we also init at domReady
		docHijack('getElementsByTagName');
		docHijack('getElementById');
		docHijack('createElement');
		addListen(doc.all);	
	}
})(window, document);

Awesome! If you are eager to add an overhead, you might be persuaded to look into http://www.w3.org/TR/DOM-Level-2-Events/events.html#Events-EventListener which allows an object instead of a function to pass as event handler. I find this a lot more maintainable, though not easily emulated in IE.

Awesome! If you are eager to add an overhead, you might be persuaded to look into http://www.w3.org/TR/DOM-Level-2-Events/events.html#Events-EventListener which allows an object instead of a function to pass as event handler. I find this a lot more maintainable, though not easily emulated in IE.

It wouldn't take much to add removeEventListener and polyfill target / currentTarget. Also, for IE<8 what if we used css expressions to do one-time check for addEventListener on an element?

Anyway, here's the concept: https://gist.github.com/2869388

It wouldn't take much to add removeEventListener and polyfill target / currentTarget. Also, for IE<8 what if we used css expressions to do one-time check for addEventListener on an element?

Anyway, here's the concept: https://gist.github.com/2869388

That is highly original, to say the least. I can suggest mouseenter and mouseleave emulation for non-IE browsers to make this an all time favorite Gist :)

That is highly original, to say the least. I can suggest mouseenter and mouseleave emulation for non-IE browsers to make this an all time favorite Gist :)

Hi! Check out my full DOM API implementation for IE < 8. It add support of [add/remove]EventListener, dispatchEvent and many other things to IE6,7,8 and w3c browsers. Short brief how it work in IE < 8 in wiki

Hi! Check out my full DOM API implementation for IE < 8. It add support of [add/remove]EventListener, dispatchEvent and many other things to IE6,7,8 and w3c browsers. Short brief how it work in IE < 8 in wiki

I forgot giving a demo: this simple demo works in IE6+ as it work in w3c browser, and this more complex demo not working in IE6 due lack of CSS, but work well in IE7+

I forgot giving a demo: this simple demo works in IE6+ as it work in w3c browser, and this more complex demo not working in IE6 due lack of CSS, but work well in IE7+

Wouldn't if(i = obj.length) (line 15) always assert true?

Wouldn't if(i = obj.length) (line 15) always assert true?

@masien In JS an assignment is an expression that returns the assigned value. So:
(x = 1) === 1 and (x = 0) === 0
So writing

if (i = obj.length) { }

is just a shortcut for:

if (obj.length) { i = obj.length; }

@masien In JS an assignment is an expression that returns the assigned value. So:
(x = 1) === 1 and (x = 0) === 0
So writing

if (i = obj.length) { }

is just a shortcut for:

if (obj.length) { i = obj.length; }

@enyo – isn't it more like this?

i = obj.length;
if (i) { }

i.e. i will always be set, regardless of the value of obj.length.

@enyo – isn't it more like this?

i = obj.length;
if (i) { }

i.e. i will always be set, regardless of the value of obj.length.

Ran into an issue with this and the current version of jQuery — since jQuery checks for "window.addEventListener" to determine the type of event system, this code seems to break it in IE8 (jQuery then assumes that removeEventListener is also present). The code in question that blew up:

function detach() {
    if ( document.addEventListener ) {
        document.removeEventListener( "DOMContentLoaded", completed, false );
        window.removeEventListener( "load", completed, false );

    } else {
        document.detachEvent( "onreadystatechange", completed );
        window.detachEvent( "onload", completed );
    }
}

(To be clear, I'm not pointing out an issue with this code, just trying to save some future Googler the headache I just had! It's rare that you'd use both an addEventListener polyfill AND jQuery)

Ran into an issue with this and the current version of jQuery — since jQuery checks for "window.addEventListener" to determine the type of event system, this code seems to break it in IE8 (jQuery then assumes that removeEventListener is also present). The code in question that blew up:

function detach() {
    if ( document.addEventListener ) {
        document.removeEventListener( "DOMContentLoaded", completed, false );
        window.removeEventListener( "load", completed, false );

    } else {
        document.detachEvent( "onreadystatechange", completed );
        window.detachEvent( "onload", completed );
    }
}

(To be clear, I'm not pointing out an issue with this code, just trying to save some future Googler the headache I just had! It's rare that you'd use both an addEventListener polyfill AND jQuery)

If you care about byte saving, there's an unnecessary var keyword on line 8. Here it is updated: https://gist.github.com/Daniel-Hug/9221945

If you care about byte saving, there's an unnecessary var keyword on line 8. Here it is updated: https://gist.github.com/Daniel-Hug/9221945

@Daniel-Hug: Yeah, that jumped out at me too. It's not just unnecessary, it's misleading. Fortunately, the order in which the various things are added to the environment record for the variable environment when entering function code is very clear in the spec and seems to be well-followed by implementations.

@Daniel-Hug: Yeah, that jumped out at me too. It's not just unnecessary, it's misleading. Fortunately, the order in which the various things are added to the environment record for the variable environment when entering function code is very clear in the spec and seems to be well-followed by implementations.

Can fn be empty? that make sense? While coding I found a "Object don´t accept this action" and checking the value show a fn = {}

Can fn be empty? that make sense? While coding I found a "Object don´t accept this action" and checking the value show a fn = {}

I've updated the code.
This code supported removeEventListener.

addEventListener & removeEventListener polyfill 1.1 - Qiita

Please rate.

I've updated the code.
This code supported removeEventListener.

addEventListener & removeEventListener polyfill 1.1 - Qiita

Please rate.

@sounisi5011 Great stuff, I took the liberty to add your work in Git to give the opportunity for anyone to contribute: https://github.com/nbouvrette/eventListenerPolyfill

I started using it in one of my projects and so far no issues.

@sounisi5011 Great stuff, I took the liberty to add your work in Git to give the opportunity for anyone to contribute: https://github.com/nbouvrette/eventListenerPolyfill

I started using it in one of my projects and so far no issues.

The original version, it occurs an error while use it with jQuery in lte IE 8 after bundled via WebPack with jQuery. It fills only addEventListener. When jQuery loaded then detect eventListener via only addEventListener then jQuery try to use removeEventListener that you can see on Ready.js in jQuery 1.12.4.

@nbouvrette Could I sure to use your version for "DOMContentLoaded" also?

The original version, it occurs an error while use it with jQuery in lte IE 8 after bundled via WebPack with jQuery. It fills only addEventListener. When jQuery loaded then detect eventListener via only addEventListener then jQuery try to use removeEventListener that you can see on Ready.js in jQuery 1.12.4.

@nbouvrette Could I sure to use your version for "DOMContentLoaded" also?

The original version, it occurs an error while use it with jQuery in lte IE 8 after bundled via WebPack with jQuery. It fills only addEventListener. When jQuery loaded then detect eventListener via only addEventListener then jQuery try to use removeEventListener that you can see on

-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-

-
-
-
-
-
-
-
-
-
-
-
-
-
-

-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-

-

-
-
-
-
-
-
-
-
-
-

-
-
-
-
-
-
-
-
-

-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-

-
-
-
-
-
-
-
-
-
-
-
-

The original version, it occurs an error while use it with jQuery in lte IE 8 after bundled via WebPack with jQuery. It fills only addEventListener. When jQuery loaded then detect eventListener via only addEventListener then jQuery try to use removeEventListener that you can see on

-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-

-
-
-
-
-
-
-
-
-
-
-
-
-
-

-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-

-

-
-
-
-
-
-
-
-
-
-

-
-
-
-
-
-
-
-
-

-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-

-
-
-
-
-
-
-
-
-
-
-
-

Fine work, thank you. There may be an issue with line 11:

fn.call(self, e);

In case that fn is an empty object. When running IE11 in emulation mode for IE8, then it reports that this method is unavailable:

Das Objekt unterstützt diese Aktion nicht.

Adding a condition helps:

if ("call" in fn) fn.call(self, e);

Fine work, thank you. There may be an issue with line 11:

fn.call(self, e);

In case that fn is an empty object. When running IE11 in emulation mode for IE8, then it reports that this method is unavailable:

Das Objekt unterstützt diese Aktion nicht.

Adding a condition helps:

if ("call" in fn) fn.call(self, e);