为什么可以注入HttpClient？

本文章基于：

    org.apache.httpcomponents
    httpclient
    4.5.6

在项目中，我们经常使用到HttpClient，但是很少用到连接池。偶然在项目中看到一位大佬把我们疯狂创建的HttpClient给注入进去：

@Autowired
private HttpClient httpClient;

看到这个代码，我表示一脸懵逼，还能这样操作？

先看下声明这个Bean的地方：

@Bean
public HttpClient httpClient() {
	return HttpClientBuilder.create()
			.setConnectionManager(connectionManager())
			.setDefaultRequestConfig(requestConfig())
			.build();
}

好像和平时的创建的差不多啊。不对，好像多设置了几个地方，设置了ConnectionManager和DefaultRequestConfig。再看一下connectionManager()和requestConfig()方法。

@Bean
public RequestConfig requestConfig() {
	RequestConfig.Builder requestConfigBuilder = RequestConfig.custom();
	//指从连接池获取连接的超时时间
	requestConfigBuilder.setConnectionRequestTimeout(httpClientProperties.getRequestConfig().getConnectionRequestTimeout());
	//指客户端和服务器建立连接的超时时间
	requestConfigBuilder.setConnectTimeout(httpClientProperties.getRequestConfig().getConnectTimeout());
	//指客户端从服务器下载数据（或页面）时，两个数据包之间的最大时间间隔
	requestConfigBuilder.setSocketTimeout(httpClientProperties.getRequestConfig().getSocketTimeout());
	return requestConfigBuilder.build();
}

@Bean
public HttpClientConnectionManager connectionManager() {
	PoolingHttpClientConnectionManager httpClientConnectionManager = new PoolingHttpClientConnectionManager();
	//连接池最大容量（即全部路由的总连接数）
	httpClientConnectionManager.setMaxTotal(httpClientProperties.getConnectionManager().getMaxTotal());
	//每个路由的默认最大连接数
	httpClientConnectionManager.setDefaultMaxPerRoute(httpClientProperties.getConnectionManager().getDefaultMaxPerRoute());
	//如果对某个站点的连接数需求比较大，可以单独设置其最大连接数
	//httpClientConnectionManager.setMaxPerRoute();
	return httpClientConnectionManager;
}

看着好像很复杂的样子。

平时创建一个HttpClient我们一般都这么写

HttpClient httpClient = HttpClients.createDefault();

或者这么写：

HttpClient httpClient = HttpClientBuilder.create().build();

其实这两种写法都没错，HttpClients.createDefault()调用的就是HttpClientBuilder.create().build()。让我们看一下HttpClientBuilder.create().build()方法：

public CloseableHttpClient build() {
	//省略部分代码......
	//如果没有连接池，则创建默认的连接池
	HttpClientConnectionManager connManagerCopy = this.connManager;
	if (connManagerCopy == null) {
		//省略部分代码......
		final PoolingHttpClientConnectionManager poolingmgr = new PoolingHttpClientConnectionManager(
				RegistryBuilder.create()
					.register("http", PlainConnectionSocketFactory.getSocketFactory())
					.register("https", sslSocketFactoryCopy)
					.build(),
				null,
				null,
				dnsResolver,
				connTimeToLive,
				connTimeToLiveTimeUnit != null ? connTimeToLiveTimeUnit : TimeUnit.MILLISECONDS);
		//省略部分代码......
		connManagerCopy = poolingmgr;
	}
	//其实就是创建MainClientExec实例的execChain
	ClientExecChain execChain = createMainExec(
					requestExecCopy,
					connManagerCopy,
					reuseStrategyCopy,
					keepAliveStrategyCopy,
					new ImmutableHttpProcessor(new RequestTargetHost(), new RequestUserAgent(userAgentCopy)),
					targetAuthStrategyCopy,
					proxyAuthStrategyCopy,
					userTokenHandlerCopy);
	//省略部分代码......
	HttpRoutePlanner routePlannerCopy = this.routePlanner;
	if (routePlannerCopy == null) {
		SchemePortResolver schemePortResolverCopy = this.schemePortResolver;
		if (schemePortResolverCopy == null) {
			schemePortResolverCopy = DefaultSchemePortResolver.INSTANCE;
		}
		if (proxy != null) {
			routePlannerCopy = new DefaultProxyRoutePlanner(proxy, schemePortResolverCopy);
		} else if (systemProperties) {
			routePlannerCopy = new SystemDefaultRoutePlanner(
					schemePortResolverCopy, ProxySelector.getDefault());
		} else {
			routePlannerCopy = new DefaultRoutePlanner(schemePortResolverCopy);
		}
	}
	//省略部分代码......
	return new InternalHttpClient(
			execChain,
			connManagerCopy,
			routePlannerCopy,
			cookieSpecRegistryCopy,
			authSchemeRegistryCopy,
			defaultCookieStore,
			defaultCredentialsProvider,
			defaultRequestConfig != null ? defaultRequestConfig : RequestConfig.DEFAULT,
			closeablesCopy);
}

它创建了一个InternalHttpClient对象，它有何不同呢？

让我们看一下InternalHttpClient产生请求的实际执行方法：

protected CloseableHttpResponse doExecute(
		final HttpHost target,
		final HttpRequest request,
		final HttpContext context) throws IOException, ClientProtocolException {
	Args.notNull(request, "HTTP request");
	HttpExecutionAware execAware = null;
	if (request instanceof HttpExecutionAware) {
		execAware = (HttpExecutionAware) request;
	}
	try {
		//省略部分代码......
		final HttpRoute route = determineRoute(target, wrapper, localcontext);
		return this.execChain.execute(route, wrapper, localcontext, execAware);
	} catch (final HttpException httpException) {
		throw new ClientProtocolException(httpException);
	}
}

在真正执行前，它执行了determineRoute()方法，根据的请求url得到一个HttpRoute：

private HttpRoute determineRoute(
		final HttpHost target,
		final HttpRequest request,
		final HttpContext context) throws HttpException {
	HttpHost host = target;
	if (host == null) {
		host = (HttpHost) request.getParams().getParameter(ClientPNames.DEFAULT_HOST);
	}
	return this.routePlanner.determineRoute(host, request, context);
}

再看下真正执行的方法MainClientExec的execute()方法（在前面说了获取到的execChain是MainClientExec）

public CloseableHttpResponse execute(
		final HttpRoute route,
		final HttpRequestWrapper request,
		final HttpClientContext context,
		final HttpExecutionAware execAware) throws IOException, HttpException {
	//省略部分代码......
	//根据HttpRoute从连接池中获取连接
	final ConnectionRequest connRequest = connManager.requestConnection(route, userToken);
	if (execAware != null) {
		if (execAware.isAborted()) {
			connRequest.cancel();
			throw new RequestAbortedException("Request aborted");
		} else {
			execAware.setCancellable(connRequest);
		}
	}

	final RequestConfig config = context.getRequestConfig();

	final HttpClientConnection managedConn;
	try {
		final int timeout = config.getConnectionRequestTimeout();
		managedConn = connRequest.get(timeout > 0 ? timeout : 0, TimeUnit.MILLISECONDS);
	} catch(final InterruptedException interrupted) {
		Thread.currentThread().interrupt();
		throw new RequestAbortedException("Request aborted", interrupted);
	} catch(final ExecutionException ex) {
		Throwable cause = ex.getCause();
		if (cause == null) {
			cause = ex;
		}
		throw new RequestAbortedException("Request execution failed", cause);
	}
	//省略部分代码......
	final ConnectionHolder connHolder = new ConnectionHolder(this.log, this.connManager, managedConn);
	try {
		//省略部分代码......
		HttpResponse response;
		for (int execCount = 1;; execCount++) {
			//省略部分代码......
			response = requestExecutor.execute(request, managedConn, context);
			//省略部分代码......
		}
		//省略部分代码......
	}//省略部分代码......
}

总结

看了以上源码，我得知PoolingHttpClientConnectionManager连接池的大概原理就是，它里面有多个路由，每个路由又有多个连接。我们每次访问都是现根据HttpHost找到HttpRoute，再根据HttpRoute获取连接。

现在我明白了为啥可以使用单例的HttpClient（实现类InternalHttpClient），它内部已经有连接池了，并不需要担心并发的时候，会出现等待或者其它问题。

如果我们平时总是有多个HttpClient，在build一个InternalHttpClient时，如果你没有设置连接池，它也会给你构建一个默认的连接池。这样就会多创建很多连接池，这就造成了浪费。