This time I will show you how Vue optimizes the project agent. What are the precautions for Vue to optimize the project agent? The following is a practical case, let's take a look.
The project structure in Vue class project development is basically similar to the Vue-cli generation method. In this method of development, the most commonly used mode is to enable the agent for mock debugging or remote debugging, that is, use the configuration proxyTable set by Vue-cli or directly use the proxy option provided by Webpack-dev-server. It uses the http-proxy library, so the specific configuration can be viewed:
https://github.com/nodejitsu/node-http-proxy#options
Using these configured parameters, we can make more flexible configurations and achieve better results
Usage requirements
Assume that our local development is currently The following states:
Local development, the data uses the local mock Server
The permissions-related interface uses local mock data, and all others use the specified One remote machine
The permission interface uses local mock data, and other data sub-interfaces use different remote machines
All interfaces use the same Remote machine
Solution
Let’s first look at the classic proxyTable writing method:
proxyTable: {
'/authui/': {
target: target,
changeOrigin: true
},
'/vendor/': {
target: target,
changeOrigin: true
}
}Copy after login
Among them The changeOrigin field is used, mainly for changing the header of the request. Refining the requirements:
Local development: The target can point to a certain port of localhost. As for the verification of the host, it is definitely not necessary.
Part of the local machine, and the other fixed remote machine: localhost and remote addresses need to be configured. Most of the remote addresses need to be verified by the host.
Same as two, but there are multiple machines: multiple machines need to be configured manually
The same remote machine, at this time the machine may need to be strictly verified, That is, the IP must also use a domain name, and the system host can be used only after configuring the system host.
Note: The difference between strict verification host and ordinary verification host is mainly that in strict verification, the requested url must be from the remote machine. Domain name,
cannot directly modify the host implementation of the requested header, that is, the domain name must be configured at the system host level.
After analyzing the specific requirements, start preparing the implementation method. The original development method is to execute npm run dev. If we need to add configuration at the command line level, we need to set it to npm run dev --param=paramvalue. For commands executed using npm's script script,
its parameters cannot be obtained through process.env, and are obtained through process.env.npm_config_paramName ,
It is not very convenient to use the ready-made command line parameter parsing library, but to save trouble, for the time being, I will use the parsing that comes with npm.
The following parameters are required during the request initiation process:
host: The host that needs to be pointed to when initiating the request may not be the same for each machine.
port: The port forwarded by the proxy
receiver: the remote address used for push, including the ip address. To save trouble, the ip address is not listed separately
Then define the proxy request custom type for configuration:
local: local address, that is, localhost
remote: The specified remote machine
Other custom types: used for other types that have been specified in the config file
Original version of the request, such as 'http://xxx' or Object type configuration, such proxy will never process
As needed, we add the following Parameters are used to control the proxy pointing address:
rd: the address of the remote machine
focus: strict mode, all custom types of proxies Convert to the specified rd machine, only available when the rd parameter exists
allLocal: Custom type proxies all point to the local
host：请求发现是否使用 host，而不是 IP 地址
总结一下（序号指向前面的需求）：
remote 类型
这么一看，貌似 host 是不需要的，它的存在主要是针对某些 机器可能需要使用 host 的方式，所以还是保留一下。
实现
逻辑理清了就很简单了，配置文件设置为：
module.export = {
rd1: {
host: 'dev1.example.com',
port: 8838,
receiver: 'http://1.1.1.1:8888/receiver'
},
rd2: {
host: 'dev2.example.com',
port: 8838,
receiver: 'http://1.1.1.1:8888/receiver'
}
}Copy after login
proxyTable 配置方式
{
proxyTable: {
'/api1': 'remote',
'/api2': 'rd2',
'/auth/xx': 'local',
'/other': 'http://example.com'
}
}Copy after login
获取 proxyTable 的代码：
// 处理 proxyTable
const releaseConfig = require('../config/release.conf.js')
const rdConfig = releaseConfig[process.env.npm_config_rd]
const isAllRemote = process.env.npm_config_focus
const useHost = isAllRemote || process.env.npm_config_host
// 是否本机开发，本机开发 remote 会指向 local
const isAllLocal = process.env.npm_config_allLocal
module.exports = function (proxy) {
const localUrl = `http://localhost:${proxy.localProxyPort}`
const defaultHost = proxy.defaultRdHost || 'dev-example.com'
const localProxyPort = proxy.localProxyPort || 8787
const finalConfig = formatReleaseConfig(releaseConfig)
const remote = finalConfig.remote || {}
if (process.env.npm_config_rd) {
if (!rdConfig) {
throw new TypeError('RD 机器名称不存在，请在 config/release.conf.js 中进行配置')
}
if (!remote.ip) {
throw new Error('请配置 rd 机器的 receiver')
}
}
if (isAllRemote && !rdConfig) {
throw new TypeError('focus 只能在提供了 rd 名称后可设置')
}
function formatReleaseConfig (config) {
const result = {}
Object.keys(config).map((key) => {
const value = config[key]
const ipMatch = (value.receiver || '').match(/:\/\/(.*?):\d/)
const ip = ipMatch && ipMatch[1]
result[key] = {
ip,
host: value.host || defaultHost,
port: value.port || '8391'
}
})
// 设置 remote
if (rdConfig) {
const ipMatch = (rdConfig.receiver || '').match(/:\/\/(.*?):\d/)
const ip = ipMatch && ipMatch[1]
result.remote = {
ip,
host: rdConfig.host || defaultHost,
port: rdConfig.port || '8391'
}
}
// 设置 local
result.local = {
ip: 'localhost',
host: 'localhost',
port: localProxyPort
}
return result
}
function setProxy (proxyTable) {
const result = {}
Object.keys(proxyTable).forEach((api) => {
let type = proxyTable[api]
const isCustomType = typeof type === 'string' && !/^http/.test(type)
if (isCustomType && type !== 'remote' && type !== 'local' && !finalConfig[type]) {
throw new TypeError(`代理类型${type}不正确，请提供 http 或 https 类型的接口，或者指定正确的 release 机器名称`)
}
if (type === 'remote' && !finalConfig.remote) {
type = 'local'
}
if (isCustomType) {
if (isAllRemote && type !== 'remote') {
type = 'remote'
}
if (isAllLocal && type !== 'local') {
type = 'local'
}
}
const targetConfig = finalConfig[type]
let target = type
if (targetConfig) {
target = {
target: `http://${useHost ? targetConfig.host : targetConfig.ip}:${targetConfig.port}`,
// 使用 host 时需要转换，其他不需要转换
headers: {
host: `${targetConfig.host}:${targetConfig.port}`
}
}
}
result[api] = target
})
return result
}
return {
proxyTable: setProxy(proxy.proxyTable),
host: remote.host || defaultHost
}
}Copy after login
用法
用法中需要配置两种指向：系统 host 和浏览器代理 Host。
之所以要两种 host, 本质上是因为接口使用的域名
和我们的本地访问的域名是相同的，同一域名无法指向两个地址，所以相当于对浏览器端进行了拦截。
系统 host 推荐使用 switchHost 进行切换，浏览器推荐使用 whistle 进行切换。
本地开发
host 配置：无
whistle 配置：默认的域名
127.0.0.1 dev.example.com
启动命令：
npm run dev
npm run dev --allLocal
Copy after login
注： 此时 proxyTable 中配置的 remote 全部转换为 local，在 allLocal 参数时将所有自定义类型转换为 local
本地 + 1 台远程
host 配置：无
whistle 配置：默认的域名
127.0.0.1 dev1.example.com
127.0.0.1 dev2.example.com
Copy after login
启动命令：
npm run dev --rd=rd1
npm run dev --rd=rd1 --host
Copy after login
注： --host 表示使用访问使用 host 而非 ip，使用时需要 host 地址
本地 + n 台远程
host 配置：无
whistle 配置：默认的域名
127.0.0.1 dev1.example.com
127.0.0.1 dev2.example.com
{
proxyTable: {
'/api1': 'rd1',
'/api2': 'rd2',
'/auth/xx': 'local',
'/other': 'http://example.com'
}
}Copy after login
proxyTable 配置：
启动命令：
npm run dev
Copy after login
远程 1 台机器
host 配置：
1.1.1.1 dev1.example.com
1.1.1.1 dev2.example.com
Copy after login
whistle 配置：默认的域名
127.0.0.1 dev1.example.com
127.0.0.1 dev2.example.com
Copy after login
启动命令：
npm run dev --rd=rd1 --focus
Copy after login
组件优化
vue 的组件化深受大家喜爱，到底组件拆到什么程度算是合理，还要因项目大小而异，小型项目可以简单几个组件搞定，甚至不用 vuex，axios 等等，如果规模较大就要细分组件，越细越好，包括布局的封装，按钮，表单，提示框，轮播等，推荐看下 Element 组件库的代码，没时间写这么详细可以直接用 Element 库，分几点进行优化
•组件有明确含义，只处理类似的业务。复用性越高越好，配置性越强越好。
•自己封装组件还是遵循配置 props 细化的规则。
•组件分类，我习惯性的按照三类划分，page、page-item 和 layout，page 是路由控制的部分，page-item 属于 page 里各个布局块如 banner、side 等等，layout 里放置多个页面至少出现两次的组件，如 icon, scrollTop 等
总结
细挖需求，可能还有更简单的方式，在大部分情况下能够减少代码修改，是 webpack 配置型的实现吧。当然，方式并不完美，尤其在 mac 下，居然不能支持 --rd xx 这种形式，可以有类似的库吧，后续可以做为深入的内容。
相信看了本文案例你已经掌握了方法，更多精彩请关注php中文网其它相关文章！
推荐阅读：
js中el表达式怎样做出非空判断
VUE对ElTableColumn进行扩展
The above is the detailed content of How Vue optimizes project proxies. For more information, please follow other related articles on the PHP Chinese website!
![[Web front-end] Node.js quick start](https://img.php.cn/upload/course/000/000/067/662b5d34ba7c0227.png)
