Spring - RestTemplate执行原理分析

什么是RestTemplate

Synchronous client to perform HTTP requests, exposing a simple, template method API over underlying HTTP client libraries such as the JDK HttpURLConnection, Apache HttpComponents, and others.

这是RestTemplate源码里对其自身的解释，从类名来看把类想设计成一个标准的模板， 简单来说就是简化Http的请求以及响应的封装，并且执行了Restful原则。如果没有RestTemplate，我们可能使用更多的还是Apache HttpClient工具。 另外从这个定义里我看到一个很重要的类HttpUrlConnection，这是RestTemplate与HTTP服务器通信的核心类。

RestTempate类结构

从类库中可以看出，这个类是一个很Spring的设计，继承抽象类InterceptingHttpAccessor，实现接口RestOperations。

RestTemplate执行流程图

开始分析之前，我们以RestTemplate#getForObject()为例，先浏览一下整个方法调用过程的执行流程图，几个关键步骤，看到其中的几个关键类，我会在后面详细说明这几个关键类。

RestTemplate构造函数

在看具体的执行http请求方法前，我们先看一下构造函数都做了那些工具。

public RestTemplate() {
    #设置各种messageConvert, 比如我们最常见的StringHttpMessageConverter。
    this.messageConverters.add(new ByteArrayHttpMessageConverter());
    this.messageConverters.add(new StringHttpMessageConverter());
    this.messageConverters.add(new ResourceHttpMessageConverter(false));
    try {
        this.messageConverters.add(new SourceHttpMessageConverter<>());
    }
    catch (Error err) {
        // Ignore when no TransformerFactory implementation is available
    }
    this.messageConverters.add(new AllEncompassingFormHttpMessageConverter());

    if (romePresent) {
        this.messageConverters.add(new AtomFeedHttpMessageConverter());
        this.messageConverters.add(new RssChannelHttpMessageConverter());
    }

    if (jackson2XmlPresent) {
        this.messageConverters.add(new MappingJackson2XmlHttpMessageConverter());
    }
    else if (jaxb2Present) {
        this.messageConverters.add(new Jaxb2RootElementHttpMessageConverter());
    }

    if (jackson2Present) {
        this.messageConverters.add(new MappingJackson2HttpMessageConverter());
    }
    else if (gsonPresent) {
        this.messageConverters.add(new GsonHttpMessageConverter());
    }
    else if (jsonbPresent) {
        this.messageConverters.add(new JsonbHttpMessageConverter());
    }

    if (jackson2SmilePresent) {
        this.messageConverters.add(new MappingJackson2SmileHttpMessageConverter());
    }
    if (jackson2CborPresent) {
        this.messageConverters.add(new MappingJackson2CborHttpMessageConverter());
    }
    #设置UriTemplateHandler
    this.uriTemplateHandler = initUriTemplateHandler();
}

构造方法总结来说只做了两件事，添加HttpMessageConvert实现类，手动配置SpringMVC的时代，想必大家都知道HttpMessageConverter吧，顾名思义就是转换HTTP请求响应过程中的消息数据。第二就是初始化UriTemplateHandler，在initUriTemplateHanlder()方法中可以看到实际实例化的是DefaultUriBuilderFactory类并返回。

RestTemplat#getForObject()

我们现在开始沿着getForObject入口来分析一下执行一个HTTP Rest请求的流程到底是怎样的哈。

getForObject

#RestTemplate.getForObject
@Override
@Nullable
public <T> T getForObject(String url, Class<T> responseType, Object... uriVariables) throws RestClientException {
   RequestCallback requestCallback = acceptHeaderRequestCallback(responseType);
   HttpMessageConverterExtractor<T> responseExtractor =
		new HttpMessageConverterExtractor<>(responseType, getMessageConverters(), logger);
   #这里传入requestCallback和responseExtractor，调用execute()
   return execute(url, HttpMethod.GET, requestCallback, responseExtractor, uriVariables);
}

execute()

#RestTemplate.execute
@Override
@Nullable
public <T> T execute(String url, HttpMethod method, @Nullable RequestCallback requestCallback,
			@Nullable ResponseExtractor<T> responseExtractor, Object... uriVariables) throws RestClientException {

  URI expanded = getUriTemplateHandler().expand(url, uriVariables);
  #调用doExecute()
  return doExecute(expanded, method, requestCallback, responseExtractor);
}

doExecute()

@Nullable
protected <T> T doExecute(URI url, @Nullable HttpMethod method, @Nullable RequestCallback requestCallback,
			@Nullable ResponseExtractor<T> responseExtractor) throws RestClientException {

    Assert.notNull(url, "URI is required");
    Assert.notNull(method, "HttpMethod is required");
    ClientHttpResponse response = null;
    try {
        #1.生成请求
        ClientHttpRequest request = createRequest(url, method);
        if (requestCallback != null) {
            #2.设置header
            requestCallback.doWithRequest(request);
        }
        #3.执行请求
        response = request.execute();
        #4.处理响应
        handleResponse(url, method, response);
        #5.返回执行结果
        return (responseExtractor != null ? responseExtractor.extractData(response) : null);
    }
    catch (IOException ex) {
        String resource = url.toString();
        String query = url.getRawQuery();
        resource = (query != null ? resource.substring(0, resource.indexOf('?')) : resource);
        throw new ResourceAccessException("I/O error on " + method.name() +
                                          " request for \"" + resource + "\": " + ex.getMessage(), ex);
    }
    finally {
        if (response != null) {
            response.close();
        }
    }
}

HttpAccessor

doExecute()方法的第一个步骤就是创建ClientHttpRequest, 这是一个接口，那么这个方法执行完会创建那个实现类呢？ createRequest() 是父类HttpAccessor的方法，我们要先简单分析下HttpAcessor。

package org.springframework.http.client.support;

import java.io.IOException;
import java.net.URI;
import java.util.ArrayList;
import java.util.List;

import org.apache.commons.logging.Log;

import org.springframework.core.annotation.AnnotationAwareOrderComparator;
import org.springframework.http.HttpLogging;
import org.springframework.http.HttpMethod;
import org.springframework.http.client.ClientHttpRequest;
import org.springframework.http.client.ClientHttpRequestFactory;
import org.springframework.http.client.ClientHttpRequestInitializer;
import org.springframework.http.client.SimpleClientHttpRequestFactory;
import org.springframework.util.Assert;

public abstract class HttpAccessor {

    protected final Log logger = HttpLogging.forLogName(getClass());

	private ClientHttpRequestFactory requestFactory = new SimpleClientHttpRequestFactory();

	private final List<ClientHttpRequestInitializer> clientHttpRequestInitializers = new ArrayList<>();

	public void setRequestFactory(ClientHttpRequestFactory requestFactory) {
		Assert.notNull(requestFactory, "ClientHttpRequestFactory must not be null");
		this.requestFactory = requestFactory;
	}

	public ClientHttpRequestFactory getRequestFactory() {
		return this.requestFactory;
	}

	public void setClientHttpRequestInitializers(
			List<ClientHttpRequestInitializer> clientHttpRequestInitializers) {

		if (this.clientHttpRequestInitializers != clientHttpRequestInitializers) {
			this.clientHttpRequestInitializers.clear();
			this.clientHttpRequestInitializers.addAll(clientHttpRequestInitializers);
			AnnotationAwareOrderComparator.sort(this.clientHttpRequestInitializers);
		}
	}
    
	public List<ClientHttpRequestInitializer> getClientHttpRequestInitializers() {
		return this.clientHttpRequestInitializers;
	}

    #在这个方法里会创建ClientHttpRequest实例并且返回。
	protected ClientHttpRequest createRequest(URI url, HttpMethod method) throws IOException {
		ClientHttpRequest request = getRequestFactory().createRequest(url, method);
		initialize(request);
		if (logger.isDebugEnabled()) {
			logger.debug("HTTP " + method.name() + " " + url);
		}
		return request;
	}

	private void initialize(ClientHttpRequest request) {
		this.clientHttpRequestInitializers.forEach(initializer -> initializer.initialize(request));
	}
}

从HttpAccessor#createRequest()分析来看,会先调用getRequestFactory() 返回_SimpleClientHttpRequestFactory，然后调用SimpleClientHttpRequestFactory.createRequest()。但是这里我们忽略了一个环节，那就是**InterceptingHttpAccessor， **_这里我们需要再深吸一口气，再看看InterceptingHttpAccessor类。

SimpleClientHttpRequestFactory类结构图

InterceptingHttpAccessor

public abstract class InterceptingHttpAccessor extends HttpAccessor {
    
    #拦截器属性，可以自定义拦截器实现业务逻辑，当然玩过SpringCloud Ribbon调用，对这个属性再熟悉不过了。
	private final List<ClientHttpRequestInterceptor> interceptors = new ArrayList<>();

    #这个属性要和HttpAccessor.requestFactory呼应一下。
	@Nullable
	private volatile ClientHttpRequestFactory interceptingRequestFactory;

	public void setInterceptors(List<ClientHttpRequestInterceptor> interceptors) {
		if (this.interceptors != interceptors) {
			this.interceptors.clear();
			this.interceptors.addAll(interceptors);
			AnnotationAwareOrderComparator.sort(this.interceptors);
		}
	}

	public List<ClientHttpRequestInterceptor> getInterceptors() {
		return this.interceptors;
	}

	@Override
	public void setRequestFactory(ClientHttpRequestFactory requestFactory) {
		super.setRequestFactory(requestFactory);
		this.interceptingRequestFactory = null;
	}


	@Override
	public ClientHttpRequestFactory getRequestFactory() {
		List<ClientHttpRequestInterceptor> interceptors = getInterceptors();
		if (!CollectionUtils.isEmpty(interceptors)) {
			ClientHttpRequestFactory factory = this.interceptingRequestFactory;
			if (factory == null) {
				factory = new InterceptingClientHttpRequestFactory(super.getRequestFactory(), interceptors);
				this.interceptingRequestFactory = factory;
			}
			return factory;
		}
		else {
			return super.getRequestFactory();
		}
	}
}

这个类的关键在于重写了父类的getRequestFacotry(), 当前做过SpringCloud Ribbon远程调用，会对这个方法格外熟悉，Spring会判断拦截器属性_interceptors是否有值，如果有值则会继续判断，然后返回InterceptionClientHttpRequestFactory, 如果拦截器属性没有值，则调用父类HttpAccessor#getRequestFactory().

RestTemplate#doExecute()

通过分析抽象父类InterceptingHttpAccessor和HttpAccessor， 我们得出结果，如果没有拦截器，只是普通的RESTFUL调用，那么最终是调用SimpleClientHttpRequestFactory#createRequest().

@Override
public ClientHttpRequest createRequest(URI uri, HttpMethod httpMethod) throws IOException {
    #再回头看看Spring对RestTemplate定义里提到的，RestTemplate的核心就是使用
    #HttpURLConnection和HTTP服务器进行通信。
    HttpURLConnection connection = openConnection(uri.toURL(), this.proxy);
    prepareConnection(connection, httpMethod.name());

    if (this.bufferRequestBody) {
        return new SimpleBufferingClientHttpRequest(connection, this.outputStreaming);
    }
    else {
        return new SimpleStreamingClientHttpRequest(connection, this.chunkSize, this.outputStreaming);
    }
}

protected HttpURLConnection openConnection(URL url, @Nullable Proxy proxy) throws IOException {
	URLConnection urlConnection = (proxy != null ? url.openConnection(proxy) : url.openConnection());
    if (!HttpURLConnection.class.isInstance(urlConnection)) {
        throw new IllegalStateException("HttpURLConnection required for [" + url + "] but got: " + urlConnection);
    }
    return (HttpURLConnection) urlConnection;
}

protected void prepareConnection(HttpURLConnection connection, String httpMethod) throws IOException {
    if (this.connectTimeout >= 0) {
        connection.setConnectTimeout(this.connectTimeout);
    }
    if (this.readTimeout >= 0) {
        connection.setReadTimeout(this.readTimeout);
    }

    connection.setDoInput(true);

    if ("GET".equals(httpMethod)) {
        connection.setInstanceFollowRedirects(true);
    }
    else {
        connection.setInstanceFollowRedirects(false);
    }

    if ("POST".equals(httpMethod) || "PUT".equals(httpMethod) ||
        "PATCH".equals(httpMethod) || "DELETE".equals(httpMethod)) {
        connection.setDoOutput(true);
    }
    else {
        connection.setDoOutput(false);
    }

	connection.setRequestMethod(httpMethod);
}

从这里代码片段分析，我们可以回到上面提的问题, 在doExecute里createRequest()最终创建的实现类是SimpleBufferingClientHttpRequest。

SimpleBufferingClientHttpRequest类结构图

SimpleStreamingClientHttpRequest类结构图

requestCallback.doWithRequest()

根据自己需要实现接口RequestCallback#doWithRequest()。

request.execute()

这里也就是调用SimplezBufferingClientHttpRequest#execute()。通过分析SimplezBufferingClientHttpRequest代码，我们知道execute()定义在抽象父类AbstractClientHttpRequest里

#AbstractClientHttpRequest#execute()
@Override
public final ClientHttpResponse execute() throws IOException {
    assertNotExecuted();
    ClientHttpResponse result = executeInternal(this.headers);
    this.executed = true;
    return result;
}

分析这个类可以看出，很经典的一个设计，父类定义行为，具体实现交给子类，这里可以看出execute里继续调用executeInternal，这是一个抽象方法，交给子类实现。

AbstractBufferingClientHttpRequest#executeInternal

@Override
protected ClientHttpResponse executeInternal(HttpHeaders headers) throws IOException {
    byte[] bytes = this.bufferedOutput.toByteArray();
    if (headers.getContentLength() < 0) {
        headers.setContentLength(bytes.length);
    }
    ClientHttpResponse result = executeInternal(headers, bytes);
    this.bufferedOutput = new ByteArrayOutputStream(0);
    return result;
}

SimpleBufferingClientHttpRequest#executeInternal

@Override
protected ClientHttpResponse executeInternal(HttpHeaders headers, byte[] bufferedOutput) throws IOException {
    addHeaders(this.connection, headers);
    // JDK <1.8 doesn't support getOutputStream with HTTP DELETE
    if (getMethod() == HttpMethod.DELETE && bufferedOutput.length == 0) {
        this.connection.setDoOutput(false);
    }
    if (this.connection.getDoOutput() && this.outputStreaming) {
        this.connection.setFixedLengthStreamingMode(bufferedOutput.length);
    }
    this.connection.connect();
    if (this.connection.getDoOutput()) {
        FileCopyUtils.copy(bufferedOutput, this.connection.getOutputStream());
    }
    else {
        // Immediately trigger the request in a no-output scenario as well
        this.connection.getResponseCode();
    }
    return new SimpleClientHttpResponse(this.connection);
}

至此，我们就拿到请求的返回结果Response了，封装SimpleClientHttpResponse并返回。似乎这几个类分析下来觉得他的调用流程并没有想象的那么复杂，整个设计还是很规范。

handleResponse

RestTemplate#handleResponse

protected void handleResponse(URI url, HttpMethod method, ClientHttpResponse response) throws IOException {
    ResponseErrorHandler errorHandler = getErrorHandler();
    boolean hasError = errorHandler.hasError(response);
    if (logger.isDebugEnabled()) {
        try {
            int code = response.getRawStatusCode();
            HttpStatus status = HttpStatus.resolve(code);
            logger.debug("Response " + (status != null ? status : code));
        }
        catch (IOException ex) {
            // ignore
        }
    }
    if (hasError) {
        errorHandler.handleError(url, method, response);
    }
}

getErrorHandler获取了一个错误处理器，如果Response的状态码是错误的，那么就调用handleError处理错误并抛出异常。

responseExtractor.extractData

ResponseExtractor是一个函数式接口，主要实现类有三个。

• HeadersExtractor - RestTemplat的内部类。
• ResponseEntityResponseExtractor - RestTemplate内部类。
• HttpMessageConverterExtractor。

前两个都是内部类，我们主要看一下HttpMessageConverterExtractor。

@Override
@SuppressWarnings({"unchecked", "rawtypes", "resource"})
public T extractData(ClientHttpResponse response) throws IOException {
    MessageBodyClientHttpResponseWrapper responseWrapper = new MessageBodyClientHttpResponseWrapper(response);
    if (!responseWrapper.hasMessageBody() || responseWrapper.hasEmptyMessageBody()) {
        return null;
    }
    MediaType contentType = getContentType(responseWrapper);

    try {
        for (HttpMessageConverter<?> messageConverter : this.messageConverters) {
            if (messageConverter instanceof GenericHttpMessageConverter) {
                GenericHttpMessageConverter<?> genericMessageConverter =
                    (GenericHttpMessageConverter<?>) messageConverter;
                if (genericMessageConverter.canRead(this.responseType, null, contentType)) {
                    if (logger.isDebugEnabled()) {
                        ResolvableType resolvableType = ResolvableType.forType(this.responseType);
                        logger.debug("Reading to [" + resolvableType + "]");
                    }
                    return (T) genericMessageConverter.read(this.responseType, null, responseWrapper);
                }
            }
            if (this.responseClass != null) {
                if (messageConverter.canRead(this.responseClass, contentType)) {
                    if (logger.isDebugEnabled()) {
                        String className = this.responseClass.getName();
                        logger.debug("Reading to [" + className + "] as \"" + contentType + "\"");
                    }
                    return (T) messageConverter.read((Class) this.responseClass, responseWrapper);
                }
            }
        }
    }
    catch (IOException | HttpMessageNotReadableException ex) {
        throw new RestClientException("Error while extracting response for type [" +
                                      this.responseType + "] and content type [" + contentType + "]", ex);
    }

    throw new RestClientException("Could not extract response: no suitable HttpMessageConverter found " +
                                  "for response type [" + this.responseType + "] and content type [" + contentType + "]");
}

可以看到，extractData先将response交给responseWrapper，如果responseWrapper有消息体且非空，则进行返回消息的读取操作。
消息的读取需要借助HttpMessageConverter接口，HttpMessageConverter具有多种实现类，以完成不同格式消息的读取，相当于消息解码器或转换头。

总结

RestTemplate提供了多种便捷访问HTTP服务的方法，提高了客户端的编码效率。底层通过使用java.net包创建HTTP请求。主要使用ClientHttpRequestFactory指定不同的HTTP请求方式。