OKHttp源码解析（一）

个人博客 迁移，欢迎光临

在上一篇博客中，我们介绍了OKHttp的基本用法，这一篇我们将从源码角度来看下OKHttp是如何完成一些列的网络的操作的。
我们知道在okhttpclient同步请求和异步请求调用的接口不一样，但它们最后都是殊途同归地走到Call里面的

private Response getResponseWithInterceptorChain(boolean forWebSocket) throws IOException {
        Call.ApplicationInterceptorChain chain = new Call.ApplicationInterceptorChain(0, this.originalRequest, forWebSocket);
        return chain.proceed(this.originalRequest);
    }

首先我们来看看调用execute和enqueue是如何走到`getResponseWithInterceptorChain的。

public Response execute() throws IOException {
        synchronized(this) {
	        //执行过不能再执行
            if(this.executed) {
                throw new IllegalStateException("Already Executed");
            }

            this.executed = true;
        }

        Response var2;
        try {
	        //把这次请求加入到分发器里
            this.client.getDispatcher().executed(this);
            Response result = this.getResponseWithInterceptorChain(false);
            if(result == null) {
                throw new IOException("Canceled");
            }

            var2 = result;
        } finally {
        //
            this.client.getDispatcher().finished(this);
        }

        return var2;
    }

上面代码14行用到了一个从OkHttpClient获得的Dispatcher然后把它加入到分发器里面的队列 executedCalls中，在完成的时候会remove掉

synchronized void executed(Call call) {
        this.executedCalls.add(call);
    }

    synchronized void finished(Call call) {
        if(!this.executedCalls.remove(call)) {
            throw new AssertionError("Call wasn\'t in-flight!");
        }
    }

再来看看异步如何走到getResponseWithInterceptorChain

public void enqueue(Callback responseCallback) {
        this.enqueue(responseCallback, false);
    }

    void enqueue(Callback responseCallback, boolean forWebSocket) {
        synchronized(this) {
            if(this.executed) {
                throw new IllegalStateException("Already Executed");
            }

            this.executed = true;
        }

        this.client.getDispatcher().enqueue(new Call.AsyncCall(responseCallback, forWebSocket, null));
    }

是不是和同步很像，最后都是调用分发器的enqueue，但和同步不同的是，同步传入enqueue方法的参数是Call，异步传入的是AsyncCall，这个是什么呢，这个是Call里面的一个内部类，而且是一个继承了Runnable的内部类，我们先来看看这个execute怎么操作

synchronized void enqueue(AsyncCall call) {
		//判断当前运行的线程是否超过最大线程数，以及同一个请求是否要超过相同请求同时存在的最大数目
        if(this.runningCalls.size() < this.maxRequests && this.runningCallsForHost(call) < this.maxRequestsPerHost) {
            this.runningCalls.add(call);
            //将请求放到线程池里运行
            this.getExecutorService().execute(call);
        } else {
	        //不满足运行条件放到后备队列里
            this.readyCalls.add(call);
        }

    }

从上面代码我们看到异步请求是有条件限制的，默认最多64个请求，而同一个请求默认最多同时存在5个

private int runningCallsForHost(AsyncCall call) {
        int result = 0;
        Iterator var3 = this.runningCalls.iterator();

        while(var3.hasNext()) {
            AsyncCall c = (AsyncCall)var3.next();
            通过比较每个请求的url，一样代表同一个请求
            if(c.host().equals(call.host())) {
                ++result;
            }
        }

        return result;
    }

下面的两个参数可以通过从OKHttpClient getDispatch得到分发器，并根据分发器提提供的设置方法去修改，修改比较简单就不贴代码了 。

private int maxRequests = 64;
    private int maxRequestsPerHost = 5;

再来看看AsyncCall 的run里面的代码

 public final void run() {
        String oldName = Thread.currentThread().getName();
        Thread.currentThread().setName(this.name);

        try {
            this.execute();
        } finally {
            Thread.currentThread().setName(oldName);
        }

    }

显然AsyncCall的execute才是核心

protected void execute() {
            boolean signalledCallback = false;

            try {
                Response e = Call.this.getResponseWithInterceptorChain(this.forWebSocket);
                if(Call.this.canceled) {
                    signalledCallback = true;
//取消call调用onFailure回调	                    this.responseCallback.onFailure(Call.this.originalRequest, new IOException("Canceled"));
                } else {
                    signalledCallback = true;
	                //请求成功，回调onResponse
                    this.responseCallback.onResponse(e);
                }
            } catch (IOException var6) {
                if(signalledCallback) {
                    Internal.logger.log(Level.INFO, "Callback failure for " + Call.this.toLoggableString(), var6);
                } else {
                    this.responseCallback.onFailure(Call.this.engine.getRequest(), var6);
                }
            } finally {
	            //完成请求，调用finish，从队列中清空
                Call.this.client.getDispatcher().finished(this);
            }

        }

在代码第五行我们又看到了getResponseWithInterceptorChain。
我们继续看getResponseWithInterceptorChain里面的实现

private Response getResponseWithInterceptorChain(boolean forWebSocket) throws IOException {
        Call.ApplicationInterceptorChain chain = new Call.ApplicationInterceptorChain(0, this.originalRequest, forWebSocket);
        return chain.proceed(this.originalRequest);
    }

创建了一个ApplicationInterceptorChain ，并且第一个参数传入0，这个0是有特殊用法的，涉及到OKHttp里面的一个功能叫做拦截器，从getResponseWithInterceptorChain这个名字里其实也能看出一二。先看看proceed做了什么

public Response proceed(Request request) throws IOException {
			//先判断是否每个拦截器都有对应的处理，没有的话先继续新建ApplicationInterceptorChain ，并执行当前拦截器的intercept方法
            if(this.index < Call.this.client.interceptors().size()) {
                Call.ApplicationInterceptorChain chain = Call.this.new ApplicationInterceptorChain(this.index + 1, request, this.forWebSocket);
                return ((Interceptor)Call.this.client.interceptors().get(this.index)).intercept(chain);
            } else {
                return Call.this.getResponse(request, this.forWebSocket);
            }
        }

这里碰到一个拦截器，OKHttp增加了一个拦截器机制，先来看看官方文档对Interceptors 的解释
Interceptors are a powerful mechanism that can monitor, rewrite, and retry calls.
解释下就是拦截器可以用来转换，重试，重写请求的机制，再来看看官方文档里提供的例子
首先自定义一个拦截器用于打印一些发送信息

class LoggingInterceptor implements Interceptor {
  @Override public Response intercept(Chain chain) throws IOException {
    Request request = chain.request();

    long t1 = System.nanoTime();
    logger.info(String.format("Sending request %s on %s%n%s",
        request.url(), chain.connection(), request.headers()));

    Response response = chain.proceed(request);

    long t2 = System.nanoTime();
    logger.info(String.format("Received response for %s in %.1fms%n%s",
        response.request().url(), (t2 - t1) / 1e6d, response.headers()));

    return response;
  }
}

然后使用的时候把它添加到okhttpclient

OkHttpClient client = new OkHttpClient();
client.interceptors().add(new LoggingInterceptor());

Request request = new Request.Builder()
    .url("http://www.publicobject.com/helloworld.txt")
    .header("User-Agent", "OkHttp Example")
    .build();

Response response = client.newCall(request).execute();
response.body().close();

当我们执行到proceed，就会去判断是否有拦截器有的话先执行拦截器里的intercept，而在intercept里一般会进行一些自定义操作并且调用procced去判断是否要继续执行拦截器操作还是直接去获取网络请求，我们看看procced做了什么

public Response proceed(Request request) throws IOException {
//判断还有拦截器需要执行不，生成新的ApplicationInterceptorChain并调用它的intercept去执行用户定义的操作
            if(this.index < Call.this.client.interceptors().size()) {
                Call.ApplicationInterceptorChain chain = Call.this.new ApplicationInterceptorChain(this.index + 1, request, this.forWebSocket);
                return ((Interceptor)Call.this.client.interceptors().get(this.index)).intercept(chain);
            } else {
                return Call.this.getResponse(request, this.forWebSocket);
            }
        }

上面的例子执行结果如下

INFO: Sending request http://www.publicobject.com/helloworld.txt on null
User-Agent: OkHttp Example

INFO: Received response for https://publicobject.com/helloworld.txt in 1179.7ms
Server: nginx/1.4.6 (Ubuntu)
Content-Type: text/plain
Content-Length: 1759
Connection: keep-alive

拦截器的整个机制如下图

上图中把拦截器氛围应用拦截器和网络拦截器，其实这个取决于你在拦截器里做了哪方面的操作，比如改变了请求头部之类的就可以成为网络拦截器。

在处理完拦截器操作后，就进入到重要的getResponse方法，真正的去进行发送请求，处理请求，接收返回结果。

Response getResponse(Request request, boolean forWebSocket) throws IOException {
        RequestBody body = request.body();
        if(body != null) {
            //如果是post方式，处理一些头部信息
            Builder followUpCount = request.newBuilder();
            MediaType response = body.contentType();
            if(response != null) {
                followUpCount.header("Content-Type", response.toString());
            }

            long followUp = body.contentLength();
            if(followUp != -1L) {
                followUpCount.header("Content-Length", Long.toString(followUp));
                followUpCount.removeHeader("Transfer-Encoding");
            } else {
                followUpCount.header("Transfer-Encoding", "chunked");
                followUpCount.removeHeader("Content-Length");
            }
            request = followUpCount.build();
        }
		//新建HttpEngine，用于进行发送请求和读取答复的细节处理
        this.engine = new HttpEngine(this.client, request, false, false, forWebSocket, (Connection)null, (RouteSelector)null, (RetryableSink)null, (Response)null);
        int var11 = 0;

        while(!this.canceled) {
            HttpEngine var13;
            try {
			     //发送请求
                this.engine.sendRequest();
                //读取答复
                this.engine.readResponse();
            } catch (RequestException var8) {
                throw var8.getCause();
            } catch (RouteException var9) {
                var13 = this.engine.recover(var9);
                if(var13 != null) {
                    this.engine = var13;
                    continue;
                }

                throw var9.getLastConnectException();
            } catch (IOException var10) {
                var13 = this.engine.recover(var10, (Sink)null);
                if(var13 != null) {
                    this.engine = var13;
                    continue;
                }

                throw var10;
            }

            Response var12 = this.engine.getResponse();
            //得到该请求对应的后续请求，比如重定向之类的
            Request var14 = this.engine.followUpRequest();
            if(var14 == null) {
                if(!forWebSocket) {
                    this.engine.releaseConnection();
                }

                return var12;
            }

            ++var11;
            if(var11 > 20) {
                throw new ProtocolException("Too many follow-up requests: " + var11);
            }

            if(!this.engine.sameConnection(var14.url())) {
                this.engine.releaseConnection();
            }

            Connection connection = this.engine.close();
            this.engine = new HttpEngine(this.client, var14, false, false, forWebSocket, connection, (RouteSelector)null, (RetryableSink)null, var12);
        }

        this.engine.releaseConnection();
        throw new IOException("Canceled");
    }

可以看到如果是post请求，先做一定的头部处理，然后新建一个HttpEngine去处理具体的操作，通过sendRequest发送具体请求操作，readResponse对服务器的答复做一定处理，在代码52行处getResponse得到从服务器返回的Response，讲到这里，我们整个的流程大概疏通了，代码贴了很多，简单的可以用下面一张图概括

整体看就是根据用户的请求Request放入到指定队列，并使用拦截器链的方式去执行我们的请求操作。
上图中的sendRequest，缓存处理，readReponse，介于篇幅问题（本来想一篇写完的，结果发现篇幅好长，分几篇完成吧）将在后续的文章中继续介绍。