自研一个简易版本的OkHTTP

一,背景

为了彻底搞明白okhttp原理，仿照okhttp自研一个

二，思路

业务上没发出一个request，使用AsyncCall包装起来，然后在网络分发器的作用下，执行具体的每一个Call,这些具体的Call会经过层层的拦截器，最终会调用到CallServiceInterceptor,这个对象里面有一个ConnectionPool,保存着每一个url对应的socket,最终处理的结果返回交给具体的每一个call,然后在回调到业务层

三，相关类图

四，具体代码实现

4.1 request

public class Request {private Map<String, String> headers;private String method;private HttpUrl url;private RequestBody body;public Request(Builder builder) {this.url = builder.url;this.method = builder.method;this.headers = builder.headers;this.body = builder.body;}public String method() {return method;}public HttpUrl url() {return url;}public RequestBody body() {return body;}public Map<String, String> headers() {return headers;}public final static class Builder {HttpUrl url;Map<String, String> headers = new HashMap<>();String method;RequestBody body;public Builder url(String url) {try {this.url = new HttpUrl(url);return this;} catch (MalformedURLException e) {throw new IllegalStateException("Failed Http Url", e);}}public Builder addHeader(String name, String value) {headers.put(name, value);return this;}public Builder removeHeader(String name) {headers.remove(name);return this;}public Builder get() {method = "GET";return this;}public Builder post(RequestBody body) {this.body = body;method = "POST";return this;}public Request build() {if (url == null) {throw new IllegalStateException("url == null");}if (TextUtils.isEmpty(method)) {method = "GET";}return new Request(this);}}
}

4.2 Response 

public class Response {int code;int contentLength = -1;Map<String, String> headers = new HashMap<>();String body;//保持连接boolean isKeepAlive;public Response() {}public Response(int code, int contentLength, Map<String, String> headers, String body,boolean isKeepAlive) {this.code = code;this.contentLength = contentLength;this.headers = headers;this.body = body;this.isKeepAlive = isKeepAlive;}public int getCode() {return code;}public int getContentLength() {return contentLength;}public Map<String, String> getHeaders() {return headers;}public String getBody() {return body;}public boolean isKeepAlive() {return isKeepAlive;}
}

4.3 HttpUrl

public class HttpUrl {String protocol;String host;String file;int port;public HttpUrl(String url) throws MalformedURLException {URL url1 = new URL(url);host = url1.getHost();file = url1.getFile();file = TextUtils.isEmpty(file) ? "/" : file;protocol = url1.getProtocol();port = url1.getPort();port = port == -1 ? url1.getDefaultPort() : port;}public String getProtocol() {return protocol;}public String getHost() {return host;}public String getFile() {return file;}public int getPort() {return port;}
}

4.4 Call

public class Call {Request request;HttpClient client;//是否执行过boolean executed;boolean cancel;public boolean isCancel() {return cancel;}public Request getRequest() {return request;}public Call(Request request, HttpClient client) {this.request = request;this.client = client;}public HttpClient getClient() {return client;}public void enqueue(Callback callback) {synchronized (this) {if (executed) {throw new IllegalStateException("已经执行过了，就不要执行");}executed = true;}//把任务交给调度器调度client.dispatcher().enqueue(new AsyncCall(callback));}/*** 是否取消*/public void cancel() {cancel = true;}/*** 执行网络请求的线程*/class AsyncCall implements Runnable {private Callback callback;public AsyncCall(Callback callback) {this.callback = callback;}@Overridepublic void run() {//是否回调过boolean singaledCallbacked = false;try {//执行真正的请求Response response = getResponse();if (cancel) {singaledCallbacked = true;callback.onFailure(Call.this, new IOException("客户端主动执行了cancel"));} else {singaledCallbacked = true;callback.onResponse(Call.this, response);}} catch (Exception e) {
//                e.printStackTrace();if (!singaledCallbacked) {//如果没有回调过callback.onFailure(Call.this, e);}} finally {//将这个任务从调度器移除client.dispatcher().finished(this);}}public String host() {return request.url().getHost();}}/*** 这里是重点！！！* @return*/private Response getResponse() throws Exception{//创建拦截器责任链List<Interceptor> interceptors = new ArrayList();//重试拦截器interceptors.add(new RetryInterceptor());//请求头拦截器interceptors.add(new HeaderInterceptor());//缓存拦截器interceptors.add(new CacheInterceptor());//连接拦截器interceptors.add(new ConnectionInterceptor());//通信拦截器interceptors.add(new CallServiceInterceptor());InterceptorChain chain = new InterceptorChain(interceptors, 0, this, null);return chain.process();}
}

4.5 InterceptorChain

public class InterceptorChain {List<Interceptor> interceptors;int index;Call call;HttpConnection httpConnection;public InterceptorChain(List<Interceptor> interceptors, int index, Call call, HttpConnection connection) {this.interceptors = interceptors;this.index = index;this.call = call;this.httpConnection = connection;}public Response process(HttpConnection httpConnection) throws IOException{this.httpConnection = httpConnection;return process();}public Response process() throws IOException {if (index >= interceptors.size()) throw new IOException("Interceptor China index out max length");//获得拦截器 去执行Interceptor interceptor = interceptors.get(index);InterceptorChain next = new InterceptorChain(interceptors, index + 1, call, httpConnection);Response response = interceptor.intercept(next);return response;}
}

4.6 Interceptor

public interface Interceptor {Response intercept(InterceptorChain chain) throws IOException;
}

4.7 ConnectionPool

public class ConnectionPool {private static final String TAG = "ConnectionPool";private static final boolean DEBUG = BuildConfig.DEBUG;private final long keepAlive;private boolean cleanrunning;private Deque<HttpConnection> connections = new ArrayDeque<>();public ConnectionPool() {this(1, TimeUnit.MINUTES);}public ConnectionPool(long keepAlive, TimeUnit unit) {this.keepAlive = unit.toMillis(keepAlive);}private Runnable cleanupRunable = new Runnable() {@Overridepublic void run() {while (true) {long waitDuration = cleanup(System.currentTimeMillis());if (waitDuration == -1) {return;}if (waitDuration > 0) {synchronized (ConnectionPool.this) {try {ConnectionPool.this.wait(waitDuration);} catch (InterruptedException e) {e.printStackTrace();}}}}}};private long cleanup(long currentTimeMillis) {long longgetIdleDuration = -1;synchronized (this) {Iterator<HttpConnection> iterator = connections.iterator();while (iterator.hasNext()) {HttpConnection connection = iterator.next();long idleDuration = currentTimeMillis - connection.getLastUseTime();//超过了最大允许闲置时间if (idleDuration > keepAlive) {if (DEBUG) Log.d(TAG, "ConnectionPool cleanup: " + "超出闲置时间，移除连接池");iterator.remove();connection.close();continue;}//没有超过闲置时间//记录 最长的闲置时间if (longgetIdleDuration < idleDuration) {longgetIdleDuration = idleDuration;}}//假如keepAlive是10s//if (longgetIdleDuration >= 0) {return keepAlive - longgetIdleDuration;}//连接池中没有连接cleanrunning = false;return longgetIdleDuration;}}private static final Executor executer = new ThreadPoolExecutor(0, Integer.MAX_VALUE,60, TimeUnit.SECONDS, new SynchronousQueue<Runnable>(), new ThreadFactory() {@Overridepublic Thread newThread(Runnable r) {Thread thread = new Thread(r, "Connection Pool");thread.setDaemon(true);//设置为守护线程,可以理解为跟进程同样的生命周期return thread;}});public void put(HttpConnection httpConnection) {//如果没有执行清理线程if (!cleanrunning) {cleanrunning = true;executer.execute(cleanupRunable);}connections.add(httpConnection);}public synchronized HttpConnection get(String host, int port) {Iterator<HttpConnection> iterator = connections.iterator();while (iterator.hasNext()) {HttpConnection connection = iterator.next();//如果查找到连接池始终在相同的host和portif (connection.isSameAddress(host, port)) {iterator.remove();return connection;}}return null;}}

4.8  CallServiceInterceptor

public class CallServiceInterceptor implements Interceptor {private static final String TAG = "CallServiceInterceptor";private static final boolean DEBUG = BuildConfig.DEBUG;@Overridepublic Response intercept(InterceptorChain chain) throws IOException {if (DEBUG) Log.d(TAG, "CallServiceInterceptor intercept: " + "通信拦截器");HttpConnection connection = chain.httpConnection;HttpCode httpCode = new HttpCode();InputStream inputStream = connection.call(httpCode);String statusLine = httpCode.readLine(inputStream);Map<String, String> headers = httpCode.readHeaders(inputStream);int contentLength = -1;if (headers.containsKey("Content-Length")) {//如果有content-length，代表可以拿到响应体的字节长度contentLength = Integer.valueOf(headers.get("Content-Length"));}boolean isChunked = false;if (headers.containsKey("Transfer-Encoding")) {//如果有有Transfer-Encoding，表示是分块编码，此时没有响应体的长度isChunked = headers.get("Transfer-Encoding").equalsIgnoreCase("chunked");}String body = null;if (contentLength > 0) {byte[] bytes = httpCode.readBytes(inputStream, contentLength);body = new String(bytes);} else if (isChunked) {body = httpCode.readChunked(inputStream);}String[] status = statusLine.split(" ");boolean isKeepAlive = false;if (headers.containsKey("Connection")) {isKeepAlive = headers.get("Connection").equalsIgnoreCase("Keep-Alive");}connection.updateLastUseTime();return new Response(Integer.valueOf(status[1]), contentLength, headers, body,isKeepAlive);}
}

本文来自互联网用户投稿，该文观点仅代表作者本人，不代表本站立场。本站仅提供信息存储空间服务，不拥有所有权，不承担相关法律责任。如若转载，请注明出处：http://www.hqwc.cn/news/111576.html

如若内容造成侵权/违法违规/事实不符，请联系编程知识网进行投诉反馈email:809451989@qq.com，一经查实，立即删除！

相关文章

【JavaEE】多线程案例-阻塞队列

【JavaEE】多线程案例-阻塞队列

1. 前言 阻塞队列&#xff08;BlockingQueue&#xff09;是一个支持两个附加操作的队列。这两个附加的操作是&#xff1a; 在队列为空时&#xff0c;获取元素的线程会等待队列变为非空当队列满时&#xff0c;存储元素的线程会等待队列可用 阻塞队列常用于生产者和消费者的场…
阅读更多...
精益求精：Android应用体积优化的终极指南

精益求精：Android应用体积优化的终极指南

精益求精&#xff1a;Android应用体积优化的终极指南 1. 介绍 在当今移动应用生态系统中&#xff0c;Android应用的体积优化是开发者需要高度重视的关键方面之一。一个庞大的应用体积不仅会对用户体验造成负面影响&#xff0c;还会导致以下问题&#xff1a; 下载速度延迟&am…
阅读更多...
Re-Learn Linux Part1

Re-Learn Linux Part1

1. Linux的目录结构 在Linux文件系统中有两个特殊的目录&#xff1a; 一个用户所在的工作目录&#xff0c;也叫当前目录&#xff0c;可以使用一个点 . 来表示&#xff1b;另一个是当前目录的上一级目录&#xff0c;也叫父目录&#xff0c;可以使用两个点 .. 来表示。 . &#…
阅读更多...
Linux 忘记root密码解决方法(CentOS7.9)

Linux 忘记root密码解决方法(CentOS7.9)

忘记Linux系统的root密码&#xff0c;可以不用重新安装系统&#xff0c;进入单用户模式重新更改一下root密码即可。 步骤如下&#xff1a; 首先&#xff0c;启动Linux系统&#xff0c;进入开机界面&#xff0c;在界面中按"e"进入编辑界面&#xff0c;动作要快。按&q…
阅读更多...
P2P协议的传输艺术

P2P协议的传输艺术

TP 采用两个 TCP 连接来传输一个文件。 控制连接&#xff1a;服务器以被动的方式&#xff0c;打开众所周知用于 FTP 的端口 21&#xff0c;客户端则主动发起连接。该连接将命令从客户端传给服务器&#xff0c;并传回服务器的应答。常用的命令有&#xff1a;list——获取文件目…
阅读更多...
如何在 Excel 中进行加，减，乘，除

如何在 Excel 中进行加，减，乘，除

在本教程中&#xff0c;我们将执行基本的算术运算&#xff0c;即加法&#xff0c;减法&#xff0c;除法和乘法。 下表显示了我们将使用的数据以及预期的结果。 | **S / N** | **算术运算符** | **第一个号码** | **第二个号码** | **结果** | | 1 | 加法&#xff08;&#xff…
阅读更多...
探索Redis速度之谜

探索Redis速度之谜

Redis&#xff0c;作为一款高性能的内存数据库&#xff0c;一直以来都因其出色的速度而闻名。然而&#xff0c;Redis的速度之快究竟源自何处&#xff0c;其中隐藏着怎样的奥秘&#xff1f;在这篇博客中&#xff0c;我们将深入探索Redis速度之谜&#xff0c;揭开其快速性能背后的…
阅读更多...
华为云云耀云服务器L实例评测｜使用宝塔10分钟部署一个围猫猫小游戏

华为云云耀云服务器L实例评测｜使用宝塔10分钟部署一个围猫猫小游戏

目录 前言一、选择华为云云耀云服务器L实例的原因二、华为云云耀云服务器的优势三、快速部署一个小游戏&#xff08;1&#xff09;终端部署1、使用Termius工具连接终端2、安装Nginx3、上传打包文件 &#xff08;2&#xff09;宝塔可视化面板部署1、进入宝塔2、宝塔菜单3、上传代…
阅读更多...
千兆以太网网络层 ARP 协议的原理与 FPGA 实现

千兆以太网网络层 ARP 协议的原理与 FPGA 实现

文章目录 前言一、ARP 帧的应用场景和存在目的二、ARP 帧工作原理三、以太网 ARP 帧发包实例设计四、以太网 CRC校验代码五、以太网 ARP 帧发包测试---GMII1.模拟数据发送2.仿真模块3.仿真波形六、以太网 ARP 帧发包测试---RGMII1.顶层文件2 .仿真代码七、上板测试（RGMII）前言…
阅读更多...
Java的序列化

Java的序列化

写在前面 本文看下序列化和反序列化相关的内容。 源码 。 1&#xff1a;为什么&#xff0c;什么是序列化和反序列化 Java对象是在jvm的堆中的&#xff0c;而堆其实就是一块内存&#xff0c;如果jvm重启数据将会丢失&#xff0c;当我们希望jvm重启也不要丢失某些对象&#xff…
阅读更多...
css自学框架之图片懒加载

css自学框架之图片懒加载

首先解释一下什么叫图片懒加载。图片懒加载是一种在页面加载时&#xff0c;延迟加载图片资源的技术&#xff0c;也就是说图片资源在需要的时候才会加载&#xff0c;就是在屏幕显示范围内加载图片&#xff0c;屏幕显示范围外图片不加载。 一、关键函数 用到的关键函数&#xf…
阅读更多...
The driver has not received any packets from the server

The driver has not received any packets from the server

在测试数据迁移时遇到的错误。 目录 一、错误 二、解决 三、数据迁移测试 3.1 环境 3.2 源码及测试 3.2.1 源码 3.2.2 测试结果&#xff08;太慢&#xff09; 3.2.3 源码修改 3.2.4 异常及解决 一、错误 The driver has not received any packets from the server. 二…
阅读更多...
推荐文章
最新文章

Copyright @ 2022~2023