前提：
如今开发方式都是采用前后台分离的方式，前台采用的方式则是单页面应用开发简称SPA，这种开发模式最大的一个特点就是将有所代码打包成了一个文件，
这会导致了一个问题就是如果这个应用过大，打出来的这个文件也会越大，性能也就随之降低了，那么如何优化这个问题呢，今天就记录下懒加载策略吧。
那么什么是懒加载呢~
懒加载概念：
懒加载也就是按需加载，只要有需要才会加载的一种策略，常用于性能优化上（个人理解）通过懒加载可以避免加载不必要的资源。
工程级别：
在微服务架构里，前台也需要同步，这就需要每个模块都需要build出属于自己的js跟css，当跳转到目标模块时，我们不需要加载其他模块的文件，于是我们就
可以利用懒加载策略进行优化
利用第三方lazyload.js对目标模块进行文件读写
第三方lazyload文件
/*jslint browser: true, eqeqeq: true, bitwise: true, newcap: true, immed: true, regexp: false */ /** LazyLoad makes it easy and painless to lazily load one or more external JavaScript or CSS files on demand either during or after the rendering of a web page. Supported browsers include Firefox 2+, IE6+, Safari 3+ (including Mobile Safari), Google Chrome, and Opera 9+. Other browsers may or may not work and are not officially supported. Visit https://github.com/rgrove/lazyload/ for more info. Copyright (c) 2011 Ryan Grove <ryan@wonko.com> All rights reserved. Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the 'Software'), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. @module lazyload @class LazyLoad @static */ LazyLoad = (function (doc) { // -- Private Variables ------------------------------------------------------ // User agent and feature test information. var env, // Reference to the <head> element (populated lazily). head, // Requests currently in progress, if any. pending = {}, // Number of times we've polled to check whether a pending stylesheet has // finished loading. If this gets too high, we're probably stalled. pollCount = 0, // Queued requests. queue = { css: [], js: [] }, // Reference to the browser's list of stylesheets. styleSheets = doc.styleSheets; // -- Private Methods -------------------------------------------------------- /** Creates and returns an HTML element with the specified name and attributes. @method createNode @param {String} name element name @param {Object} attrs name/value mapping of element attributes @return {HTMLElement} @private */ function createNode(name, attrs) { var node = doc.createElement(name), attr; for (attr in attrs) { if (attrs.hasOwnProperty(attr)) { node.setAttribute(attr, attrs[attr]); } } return node; } /** Called when the current pending resource of the specified type has finished loading. Executes the associated callback (if any) and loads the next resource in the queue. @method finish @param {String} type resource type ('css' or 'js') @private */ function finish(type) { var p = pending[type], callback, urls; if (p) { callback = p.callback; urls = p.urls; urls.shift(); pollCount = 0; // If this is the last of the pending URLs, execute the callback and // start the next request in the queue (if any). if (!urls.length) { callback && callback.call(p.context, p.obj); pending[type] = null; queue[type].length && load(type); } } } /** Populates the <code>env</code> variable with user agent and feature test information. @method getEnv @private */ function getEnv() { var ua = navigator.userAgent; env = { // True if this browser supports disabling async mode on dynamically // created script nodes. See // http://wiki.whatwg.org/wiki/Dynamic_Script_Execution_Order async: doc.createElement('script').async === true }; (env.webkit = /AppleWebKit//.test(ua)) || (env.ie = /MSIE|Trident/.test(ua)) || (env.opera = /Opera/.test(ua)) || (env.gecko = /Gecko//.test(ua)) || (env.unknown = true); } /** Loads the specified resources, or the next resource of the specified type in the queue if no resources are specified. If a resource of the specified type is already being loaded, the new request will be queued until the first request has been finished. When an array of resource URLs is specified, those URLs will be loaded in parallel if it is possible to do so while preserving execution order. All browsers support parallel loading of CSS, but only Firefox and Opera support parallel loading of scripts. In other browsers, scripts will be queued and loaded one at a time to ensure correct execution order. @method load @param {String} type resource type ('css' or 'js') @param {String|Array} urls (optional) URL or array of URLs to load @param {Function} callback (optional) callback function to execute when the resource is loaded @param {Object} obj (optional) object to pass to the callback function @param {Object} context (optional) if provided, the callback function will be executed in this object's context @private */ function load(type, urls, callback, obj, context) { var _finish = function () { finish(type); }, isCSS = type === 'css', nodes = [], i, len, node, p, pendingUrls, url; env || getEnv(); if (urls) { // If urls is a string, wrap it in an array. Otherwise assume it's an // array and create a copy of it so modifications won't be made to the // original. urls = typeof urls === 'string' ? [urls] : urls.concat(); // Create a request object for each URL. If multiple URLs are specified, // the callback will only be executed after all URLs have been loaded. // // Sadly, Firefox and Opera are the only browsers capable of loading // scripts in parallel while preserving execution order. In all other // browsers, scripts must be loaded sequentially. // // All browsers respect CSS specificity based on the order of the link // elements in the DOM, regardless of the order in which the stylesheets // are actually downloaded. if (isCSS || env.async || env.gecko || env.opera) { // Load in parallel. queue[type].push({ urls: urls, callback: callback, obj: obj, context: context }); } else { // Load sequentially. for (i = 0, len = urls.length; i < len; ++i) { queue[type].push({ urls: [urls[i]], callback: i === len - 1 ? callback : null, // callback is only added to the last URL obj: obj, context: context }); } } } // If a previous load request of this type is currently in progress, we'll // wait our turn. Otherwise, grab the next item in the queue. if (pending[type] || !(p = pending[type] = queue[type].shift())) { return; } head || (head = doc.head || doc.getElementsByTagName('head')[0]); pendingUrls = p.urls.concat(); for (i = 0, len = pendingUrls.length; i < len; ++i) { url = pendingUrls[i]; if (isCSS) { node = env.gecko ? createNode('style') : createNode('link', { href: url, rel: 'stylesheet' }); } else { node = createNode('script', { src: url }); node.async = false; } node.className = 'lazyload'; node.setAttribute('charset', 'utf-8'); if (env.ie && !isCSS && 'onreadystatechange' in node && !('draggable' in node)) { node.onreadystatechange = function () { if (/loaded|complete/.test(node.readyState)) { node.onreadystatechange = null; _finish(); } }; } else if (isCSS && (env.gecko || env.webkit)) { // Gecko and WebKit don't support the onload event on link nodes. if (env.webkit) { // In WebKit, we can poll for changes to document.styleSheets to // figure out when stylesheets have loaded. p.urls[i] = node.href; // resolve relative URLs (or polling won't work) pollWebKit(); } else { // In Gecko, we can import the requested URL into a <style> node and // poll for the existence of node.sheet.cssRules. Props to Zach // Leatherman for calling my attention to this technique. node.innerHTML = '@import "' + url + '";'; pollGecko(node); } } else { node.onload = node.onerror = _finish; } nodes.push(node); } for (i = 0, len = nodes.length; i < len; ++i) { head.appendChild(nodes[i]); } } /** Begins polling to determine when the specified stylesheet has finished loading in Gecko. Polling stops when all pending stylesheets have loaded or after 10 seconds (to prevent stalls). Thanks to Zach Leatherman for calling my attention to the @import-based cross-domain technique used here, and to Oleg Slobodskoi for an earlier same-domain implementation. See Zach's blog for more details: http://www.zachleat.com/web/2010/07/29/load-css-dynamically/ @method pollGecko @param {HTMLElement} node Style node to poll. @private */ function pollGecko(node) { var hasRules; try { // We don't really need to store this value or ever refer to it again, but // if we don't store it, Closure Compiler assumes the code is useless and // removes it. hasRules = !!node.sheet.cssRules; } catch (ex) { // An exception means the stylesheet is still loading. pollCount += 1; if (pollCount < 200) { setTimeout(function () { pollGecko(node); }, 50); } else { // We've been polling for 10 seconds and nothing's happened. Stop // polling and finish the pending requests to avoid blocking further // requests. hasRules && finish('css'); } return; } // If we get here, the stylesheet has loaded. finish('css'); } /** Begins polling to determine when pending stylesheets have finished loading in WebKit. Polling stops when all pending stylesheets have loaded or after 10 seconds (to prevent stalls). @method pollWebKit @private */ function pollWebKit() { var css = pending.css, i; if (css) { i = styleSheets.length; // Look for a stylesheet matching the pending URL. while (--i >= 0) { if (styleSheets[i].href === css.urls[0]) { finish('css'); break; } } pollCount += 1; if (css) { if (pollCount < 200) { setTimeout(pollWebKit, 50); } else { // We've been polling for 10 seconds and nothing's happened, which may // indicate that the stylesheet has been removed from the document // before it had a chance to load. Stop polling and finish the pending // request to prevent blocking further requests. finish('css'); } } } } return { /** Requests the specified CSS URL or URLs and executes the specified callback (if any) when they have finished loading. If an array of URLs is specified, the stylesheets will be loaded in parallel and the callback will be executed after all stylesheets have finished loading. @method css @param {String|Array} urls CSS URL or array of CSS URLs to load @param {Function} callback (optional) callback function to execute when the specified stylesheets are loaded @param {Object} obj (optional) object to pass to the callback function @param {Object} context (optional) if provided, the callback function will be executed in this object's context @static */ css: function (urls, callback, obj, context) { load('css', urls, callback, obj, context); }, /** Requests the specified JavaScript URL or URLs and executes the specified callback (if any) when they have finished loading. If an array of URLs is specified and the browser supports it, the scripts will be loaded in parallel and the callback will be executed after all scripts have finished loading. Currently, only Firefox and Opera support parallel loading of scripts while preserving execution order. In other browsers, scripts will be queued and loaded one at a time to ensure correct execution order. @method js @param {String|Array} urls JS URL or array of JS URLs to load @param {Function} callback (optional) callback function to execute when the specified scripts are loaded @param {Object} obj (optional) object to pass to the callback function @param {Object} context (optional) if provided, the callback function will be executed in this object's context @static */ js: function (urls, callback, obj, context) { load('js', urls, callback, obj, context); } }; })(this.document);
封装路由
class LazyLoader extends React.Component { constructor(props) { super(props); this.state = { module: null } } componentWillMount() { this.load(this.props.module, this.props.path); } componentWillReceiveProps(newProps) { this.load(newProps.module, newProps.path); } load(module, path) { let cssPath = path + '', jsPath = path + ''; LazyLoad.css(cssPath); LazyLoad.js(jsPath, () => { this.state.module = Modules[module]; }) } render() { if (!this.state.module) return null; else this.props.children(this.state.module); } } const HTMLModule = ({ name, path }) => { let param = location.param; location.param = null; let ModuleComponent = Modules[name]; return ( <React.Fragment> { ModuleComponent ? <ModuleComponent param={param} /> : <LazyLoader path={path} module={name}> {(Item) => <Item param={param} />} </LazyLoader> } </React.Fragment> ); }; class ModulesRouter extends React.Component { constructor(props) { super(props); } render() { return ( <Switch> <Route path={'/Module/Target'} render={() => <HTMLModule name='Target' path={'./Module/Target'} />} /> <Redirect from='*' to='/403' /> </Switch> ); } } export default ModulesRouter;
路由与组件级别：
这里我们利用React16.6.X以后的版本新特性React.lazy跟React.Suspense进行路由以及组件部分的优化
路由设计
root路由
const RootRouter = () => { return <HashRouter> <React.Suspense fallback={<div>loading...</div>}> <Switch> <Route exact path="/" component={Home} /> <Route path="/oAuthPromisition" component={oAuthPromisition} /> <Route path="/WeChatPromisitionCheck" component={WeChatPromisitionCheck} /> <Route path="/403" component={Forbidden} /> <Route path="/sub" component={ModuleRouter} /> </Switch> </React.Suspense> </HashRouter> }
子路由设计
import { Route, Switch } from 'react-router-dom'; function lazyWithPreload(factory) { const Component = React.lazy(factory); Component.preload = factory; return Component; } const LeaveMsg = lazyWithPreload(() => { return import(/* webpackChunkName: "leavemessage" */ '../JSX/Components/LeaveMessage/LeaveMessage') }); const Blog = lazyWithPreload(() => { return import(/* webpackChunkName: "blog" */ '../JSX/Components/Blog/Blog') }); const Article = lazyWithPreload(() => { return import(/* webpackChunkName: "article" */ '../JSX/Components/Blog/Article') }); const Production = lazyWithPreload(() => { return import(/* webpackChunkName: "production" */ '../JSX/Components/Production/index') }); const About = lazyWithPreload(() => { return import(/* webpackChunkName: "about" */ '../JSX/Components/About/About') }); const Write = lazyWithPreload(() => { return import(/* webpackChunkName: "write" */ '../JSX/Components/Blog/Write') }); const Overview = lazyWithPreload(() => { return import(/* webpackChunkName: "overview" */ '../JSX/Components/Overview/Overview') }); function moduleHtml(component) { component.preload() return component; } const SubRouter = () => { return <Switch> <Route path="/sub/leavemessage" component={moduleHtml(LeaveMsg)} /> <Route path="/sub/blog" component={moduleHtml(Blog)} /> <Route path="/sub/article" component={moduleHtml(Article)} /> <Route path="/sub/production" component={moduleHtml(Production)} /> <Route path="/sub/about" component={moduleHtml(About)} /> <Route path="/sub/write" component={moduleHtml(Write)} /> <Route path="/sub/overview" component={moduleHtml(Overview)} /> </Switch> } export default SubRouter;
路由设计总结：
利用import动态加载组件，以达到按需加载的目的，但是在动态加载的时候webpack会发生代码切割，所有我们同时也要进行webpack配置
首先配置分割规则
output: { path: path.resolve(__dirname, "./public/"), filename: '[name].bundle.js', chunkFilename: '[name].module.js', //非入口文件命名规则 },
配置chunkFilename即非入口文件命名规则，就是这些由webpack分割的文件命名规范
由于webpack分割之后的路径查找方式是通过build输出路径查找的，因此如果我们单纯的将静态文件(index.html)当做根文件来加载，是找不到分割之后的代码路径的，
因此需要配置HtmlWebpackPlugin
new HtmlWebpackPlugin({ template: 'Index.html', filename: 'index.html', chunks: [ 'index' ]
}),
然后将资源文件，以及第三方文件拷贝到build输出文件夹下
new CopyWebpackPlugin([ { from: __dirname + '/Lib/editor', to: __dirname + '/public/editor', // toType: 'directory', //file 或者 dir 可选，默认是文件 // ignore: ['.*'] },{ from: __dirname + '/picSrc', to: __dirname + '/public/picSrc', } ]),
这样webpack配置修改完成了
组件级别跟路由级别基本类似，但是我们要考虑到，如果这么拆分代码，会产生很多文件，根据实际情况使用。