chromium-cronet文档 原文文档写的已经很清楚，最好还是参考官方文档，避免由于版本原因导致的问题。

Cronet开发者文档：https://developer.android.com/guide/topics/connectivity/cronet

博客中的git地址：https://github.com/bgylde/chromium-cornet

环境配置

1. chromium源码环境；
2. 已经配置好相关环境，安装好相关依赖；
3. 这里是在linux环境下对Android库的编译，在macos环境下会直接编译为ios平台库；

编译开发和debug环境的Cronet库

Android / IOS build

$ ./components/cronet/tools/cr_cronet.py gn --out_dir=out/Cronet

如果build主机是在linux环境下，build的是android的库。如果build主机是macOS，build的是ios库。

$ ninja -C out/Cronet cronet_package

编译Cronet库，最终文件可以在out/Cronet/cornet中寻找。

使用cronet库

1. 创建CronetEngine，最好整个应用使用一个CronetEngine，这里可以理解为OkHttpClient；
2. 创建自己的线程池给Cronet使用，Cronet的网络请求都会在线程池中，避免主线程阻塞；
3. 实现回调UrlRequest.Callback，在UrlRequest调用start以后，网络请求开始，之后产生请求回调；

public class UrlRequestCallback extends UrlRequest.Callback {private static final String TAG = "UrlRequestCallback";private long startTime;private ByteBuffer mByteBuffer = ByteBuffer.allocateDirect(102400);private ByteArrayOutputStream mBytesReceived = new ByteArrayOutputStream();private WritableByteChannel mReceiveChannel = Channels.newChannel(mBytesReceived);@Overridepublic void onRedirectReceived(UrlRequest request, UrlResponseInfo info, String newLocationUrl) {LogUtils.i(TAG, "onRedirectReceived method called.");// You should call the request.followRedirect() method to continue// processing the request.request.followRedirect();}@Overridepublic void onResponseStarted(UrlRequest request, UrlResponseInfo info) {LogUtils.i(TAG, "onResponseStarted method called.");// You should call the request.read() method before the request can be// further processed. The following instruction provides a ByteBuffer object// with a capacity of 102400 bytes to the read() method.//ByteBuffer byteBuffer = ByteBuffer.allocateDirect(102400);startTime = System.currentTimeMillis();request.read(mByteBuffer);}@Overridepublic void onReadCompleted(UrlRequest request, UrlResponseInfo info, ByteBuffer byteBuffer) {LogUtils.i(TAG, "onReadCompleted method called.");// You should keep reading the request until there&＃39;s no more data.try {byteBuffer.flip();mReceiveChannel.write(byteBuffer);byteBuffer.clear();} catch (IOException e) {LogUtils.e(TAG, e);}request.read(mByteBuffer);}@Overridepublic void onSucceeded(UrlRequest request, UrlResponseInfo info) {LogUtils.i(TAG, "onSucceeded method called: " + info.toString());LogUtils.d(TAG, "cost time: " + (System.currentTimeMillis() - startTime) + " ms");//byte[] bytes = mBytesReceived.toByteArray();String receivedData = null;try {receivedData = mBytesReceived.toString("GBK");} catch (UnsupportedEncodingException e) {LogUtils.e(TAG, e);}final String url = info.getUrl();final String text = "Completed " + url + " (" + info.getHttpStatusCode() + ")";//LogUtils.i(TAG, "text: " + text);//LogUtils.i(TAG, "receivedData: " + receivedData);}@Overridepublic void onFailed(UrlRequest urlRequest, UrlResponseInfo urlResponseInfo, CronetException e) {//LogUtils.d(TAG, "url: " + urlResponseInfo.getUrl());LogUtils.e(TAG, "onFailed method called.", e);}@Overridepublic void onCanceled(UrlRequest request, UrlResponseInfo info) {LogUtils.d(TAG, "onCanceled method called.");}
}

• onRedirectReceived 顾名思义，是重定向的回调，这里直接选择继续访问重定向的地址，也可以调用UrlRequest的cancel方法，取消访问；
• onResponseStarted 从google文档上面来看是请求完header以后开始请求body部分会回调这里，每次请求只会回调一次；
• onReadCompleted 这是读取body一定数据时会回调这个方法，这里request.read读取的数据不一定会填满缓冲区，请求生命周期中会有多次回调发生；
• onSucceeded 最终请求成功回调，可以作为数据处理阶段；
• onFailed 请求失败回调，例如网络不通，或者没有网络访问权限之类的错误；
• onCanceled 请求取消回调，只会在cancel后才会回调，回调这个意味着整个请求完成；

https://ssl.gstatic.com/gb/images/qi2_00ed8ca1.png 实现quic访问，可作为测试地址。

 这是通过抓包看到的quic协议，进一步的性能对比就需要自己搭建一个支持quic协议的服务器。