While I ended up using a CSS-only implementation for this pen, I started by writing it mostly using classes and JavaScript.

However, I had a conflict. I wanted to use Delegated Events but I also wanted to minimize the dependancies I wanted to inject. I didn't want to have to import all of jQuery for this little test, just to be able to use delegated events one bit.

Let's take a closer look at what exactly delegated events are, how they work, and various ways to implement them.

Ok, so what's the issue?

Let's look at a simplified example:

Let's say that I had a list of buttons and each time I clicked on one, then I want to mark that button as "active". If I click it again, then deactivate it.

So let's start with some HTML:

<ul class="toolbar">
<li><button class="btn">Pencil</button></li>
<li><button class="btn">Pen</button></li>
<li><button class="btn">Eraser</button></li>
</ul>

I could use standard JavaScript event handler by doing something like this:

var buttons = document.querySelectorAll(".toolbar .btn");

for(var i = 0; i < buttons.length; i++) {
var button = buttons[i];
button.addEventListener("click", function() {
if(!button.classList.contains("active"))
button.classList.add("active");
else
button.classList.remove("active");
});
}

And this looks good... but it wont' work... Not the way one might expect it to.

Bitten by closures

For those of you that have been doing functional JavaScript for a while, the problem is pretty obvious.

For the uninitiated, the handler function closes over the button variable. However, there is only one of them; it gets reassigned by each iteration of the loop.

The first time though the loop, it points a the first button. The next time, the second button. And so on. However, by the time that you click one of the elements, the loop has completed and the button variable will point at the last element iterated over. Not good.

What we really need is a stable scope for each function; let's refactor an extract a handler generator to give us a stable scope:

var buttons = document.querySelectorAll(".toolbar button");
var createToolbarButtonHandler = function(button) {
return function() {
if(!button.classList.contains("active"))
button.classList.add("active");
else
button.classList.remove("active");
};
}; for(var i = 0; i < buttons.length; i++) {
buttons[i].addEventListener("click", createToolBarButtonHandler(buttons[i]));
}

Better! And now it actually works. We are using a function to create us a stable scope forbutton, so the button in the handler will always point at the element that we think it will.

So, what the problem?

This seems good and it will work for the most part. However, we can still do better.

First, we are making a lot of handlers. For each element that matches .toolbar buttonwe create a function and attach it as an event listener. With the three buttons we have right now the allocations are negligible.

However, what if we had:

<ul class="toolbar">
<li><button id="button_0001">Foo</button></li>
<li><button id="button_0002">Bar</button></li>
// ... 997 more elements ...
<li><button id="button_1000">baz</button></li>
</ul>

It won't blow up, but it is far from ideal. We are allocating a bunch of function that we don't have to. Let's try to refactor so that we can share a single function that is attachedmultiple times.

Rather than closing over the button variable to keep track of which button we clicked on, we can use event object that is handed to each event handler as the first argument.

The event object contains some metadata about the event. In this case, we want the thecurrentTarget property of the event to get a reference to the element that was actually clicked on.

var buttons = document.querySelectorAll(".toolbar button");

var toolbarButtonHandler = function(e) {
var button = e.currentTarget;
if(!button.classList.contains("active"))
button.classList.add("active");
else
button.classList.remove("active");
}; for(var i = 0; i < buttons.length; i++) {
button.addEventListener("click", toolbarButtonHandler);
}

Great! This not only simplified down to a single function that is added multiple times, also made the code more readable by factoring out our generator function.

But, we can still do better.

Let's say we added some buttons dynamically into the list. Then we would also need to remember to wire up the event listeners directly to those dynamic elements. And we would have to hold onto a reference to that handler and reference from more places. That doesn't sound like fun.

Perhaps there is a different approach.

Let's start by getting a better understanding of how events work and how they move through the DOM.

Okay, how do (most) events work?

When the user clicks on an element, an event gets generated to notify the application of the user's intent. Events get dispatched in three phases:

  • Capturing
  • Target
  • Bubbling

NOTE: Not all events bubble/capture, instead they are dispatched directly on the target, but most do.

The event starts outside the document and then descends though the DOM hierarchy to the target of the event. Once the event reaches it's target, it then turns around and heads back out the same way, until it exits the DOM.

Here is a full HTML example:

<html>
<body>
<ul>
<li id="li_1"><button id="button_1">Button A</button></li>
<li id="li_2"><button id="button_2">Button B</button></li>
<li id="li_3"><button id="button_3">Button C</button></li>
</ul>
</body>
</html>

If the user clicks on Button A, then the event would travel like this like this:

START
| #document \
| HTML |
| BODY } CAPTURE PHASE
| UL |
| LI#li_1 /
| BUTTON <-- TARGET PHASE
| LI#li_1 \
| UL |
| BODY } BUBBLING PHASE
| HTML |
v #document /
END

Notice that you can follow the path the event takes down to the element that gets clicked on. For any button we click on in our DOM, we can be sure that the event will bubble back out through our parent ul element. We can exploit this feature of the event dispatcher, combined with our defined hierarchy to simplify our implementation and implement Delegated Events.

Delegated Events

Delegated events are events that are attached to a parent element, but only get executed when the target of the event matches some criteria.

Let's look at a concrete example and switch back to our toolbar example DOM from before:

<ul class="toolbar">
<li><button class="btn">Pencil</button></li>
<li><button class="btn">Pen</button></li>
<li><button class="btn">Eraser</button></li>
</ul>

So, since we know that any clicks on the button elements will get bubbled through theUL.toolbar element, let's put the event handler there instead. We'll have to adjust our handler a little bit from before;

var toolbar = document.querySelector(".toolbar");
toolbar.addEventListener("click", function(e) {
var button = e.target;
if(!button.classList.contains("active"))
button.classList.add("active");
else
button.classList.remove("active");
});

That cleaned up a lot of code, and we have no more loops! Notice that we use e.targetinstead of e.currentTarget as we did before. That is because we are listening for the event at a different level.

  • e.target is actual target of the event. Where the event is trying to get to, or where it came from, in the DOM.
  • e.currentTarget is the current element that is handling the event.

In our case e.currentTarget will be the UL.toolbar.

More Robust Delegated Events

Right now, we handle any click on any element that bubbles though UL.toolbar, but our matching strategy is a little too simple. What if we had more complicated DOM that included icons and items that were supposed to be non-clickable

<ul class="toolbar">
<li><button class="btn"><i class="fa fa-pencil"></i> Pencil</button></li>
<li><button class="btn"><i class="fa fa-paint-brush"></i> Pen</button></li>
<li class="separator"></li>
<li><button class="btn"><i class="fa fa-eraser"></i> Eraser</button></li>
</ul>

OOPS. Now, when we click on the LI.separator or the icons, we add the active class to that element. That's not cool. We need a way to filter our events so we only react to elements we care about, or if our target element is contained by an element we care about.

Let's make a little helper to handle that:

var delegate = function(criteria, listener) {
return function(e) {
var el = e.target;
do {
if (!criteria(el)) continue;
e.delegateTarget = el;
listener.apply(this, arguments);
return;
} while( (el = el.parentNode) );
};
};

This helper does two things, first it walks though each element and their parents to see if it matches a criteria function. If it does, then it adds a property to the event object calleddelegateTarget, which is the element that matched our filtering criteria. And then invokes the listener. If nothing matches, the no handlers are fired.

We can use it like this:

var toolbar = document.querySelector(".toolbar");
var buttonsFilter = function(elem) { return elem.classList && elem.classList.contains("btn"); };
var buttonHandler = function(e) {
var button = e.delegateTarget;
if(!button.classList.contains("active"))
button.classList.add("active");
else
button.classList.remove("active");
};
toolbar.addEventListener("click", delegate(buttonsFilter, buttonHandler));

BOOM! That's what I'm talking about: A single event handler, attached to a single element that does all the work, but only does it on the elements that we care about and will react nicely to elements added or removed from the DOM dynamically.

Wrapping up

We've looked at the basics of how to implement event delegation in pure javascript in order to reduce the number of event handlers we need to generate or attach.

There are a few things I would do, if I were going to abstract this into a library, or use it for production level code:

  • Create helper functions to handle criteria matching in a unified functional way. Something like:
var criteria = {
isElement: function(e) { return e instanceof HTMLElement; },
hasClass: function(cls) {
return function(e) {
return criteria.isElement(e) && e.classList.contains(cls);
}
}
// More criteria matchers
};
  • A partial application helper would also be nice:
var partialDelgate = function(criteria) {
return function(handler) {
return delgate(criteria, handler);
}
};

If you have any suggestions or improvments, drop a comment or send me a message! Happy coding!

http://codepen.io/32bitkid/blog/understanding-delegated-javascript-events

Understanding Delegated JavaScript Events的更多相关文章

  1. Understanding the JavaScript Engine—— two phase

    Understanding the JavaScript Engine — Part 1   I have been a Ruby on Rails developer for the last 2 ...

  2. Javascript Events

    事件通常与函数配合使用,这样就可以通过发生的事件来驱动函数执行. 事件句柄 html4.0的新特性之一是有能力使html事件触发浏览器中的动作action,比如当用户点击某个html元素时启动一段Ja ...

  3. Javascript Madness: Mouse Events

    http://unixpapa.com/js/mouse.html Javascript Madness: Mouse Events Jan WolterAug 12, 2011 Note: I ha ...

  4. JavaScript Interview Questions: Event Delegation and This

    David Posin helps you land that next programming position by understanding important JavaScript fund ...

  5. 推荐15款制作 SVG 动画的 JavaScript 库

    在当今时代,SVG是最流行的和正在被众多的设计人员和开发人员使用,创建支持视网膜和响应式的网页设计.绘制SVG不是一个艰巨的任务,因为大量的 JavaScript 库可与 SVG 图像搭配使用.这些J ...

  6. 常用的Javascript设计模式

    <parctical common lisp>的作者曾说,如果你需要一种模式,那一定是哪里出了问题.他所说的问题是指因为语言的天生缺陷,不得不去寻求和总结一种通用的解决方案. 不管是弱类型 ...

  7. Javascript——浅谈 Event Flow

    1.Javascript Events : Event Bubbling(事件冒泡) 如果事件从最特定的元素开始,则事件流中的一个阶段称为事件冒泡(DOM中可能最深的节点)然后向上流向最不特定的节点( ...

  8. 集成Javascript Logging on MVC or Core

    ASP.NET Core provides us a rich Logging APIs which have a set of logger providers including: Console ...

  9. 每个JavaScript工程师都应懂的33个概念

    摘要: 基础很重要啊! 原文:33 concepts every JavaScript developer should know 译文:每个 JavaScript 工程师都应懂的33个概念 作者:s ...

随机推荐

  1. 两种MVC框架比较

    基于Web的MVC framework在J2EE的世界内已是空前繁荣. TTS网站上几乎每隔一两个星期就会有新的MVC框架发布.目前比较好的MVC,老牌的有Struts.Webwork.新兴的MVC框 ...

  2. postgresql+postgis+pgrouting实现最短路径查询(2)---openlayers+geoserver实现最短路径

    自己的最短路径实现基本上是按照参考博文的1.2和3进行的,实现的时候也是问题不断,只能是一个一个解决. 问题1:自己发布的geoserver服务无法和OSM底图叠加到一起. 解决:参考博文2提到发布服 ...

  3. jquery Jbox 插件实现弹出窗口在修改的数据之后,关闭弹出窗口刷新父页面的问题

    http://blog.csdn.net/nsdnresponsibility/article/details/51282797 问题如题: 这里我们在父页面定义一个全局的变量来标识是否需要刷新父页面 ...

  4. 理解基本包装类型Number,String,Boolean

    在前面我们知道了引用类型是什么了,也就能理解包装类型了.包装对象其实也是一种引用类型,之所以要单独提出来只不过是因为它们可以把原始类型的值变成(包装成)对象,这样它们也就获得了各自类型相应的特殊行为了 ...

  5. 如何指定安装webpack

    在此我要推荐webpack简易学习教程:https://www.runoob.com/w3cnote/webpack-tutorial.html 大家可以参考这个菜鸟教程,但是这个菜鸟教程有其局限性, ...

  6. MongoDB的角色作用(2)

    数据压力大到机器支撑不了的时候能否做到自动扩展? 在系统早期,数据量还小的时候不会引起太大的问题,但是随着数据量持续增多,后续迟早会出现一台机器硬件瓶颈问题的.而MongoDB主打的就是海量数据架构, ...

  7. 智能门锁超低功耗:SI522(13.56芯片)替代MFRC522\FM17522

    SI522(超低功耗13.56M芯片)替代RC522 完全兼容 PIN对PIN,同时也替代FM17522. MF RC522 是应用于13.56MHz 非接触式通信中高集成度读写卡系列芯片中的一员.是 ...

  8. Java中线程同步的理解

    我们可以在计算机上运行各种计算机软件程序.每一个运行的程序可能包括多个独立运行的线程(Thread). 线程(Thread)是一份独立运行的程序,有自己专用的运行栈.线程有可能和其他线程共享一些资源, ...

  9. 手工检测SQL注入(安全性测试)

    手动你的ASP站可否注入: http://127.0.0.1/xx?id=11 and 1=1 (正常页面) http://127.0.0.1/xx?id=11 and 1=2 (出错页面) 检测表段 ...

  10. Oracle分区表删除分区引发错误ORA-01502: 索引或这类索引的分区处于不可用状态

    (一)问题: 最近在做Oracle数据清理,在对分区表进行数据清理时,采用的方法是drop partition,删除的过程中,没有遇到任何问题,大概过了10分钟,开发人员反馈部分分区表上的业务失败.具 ...