RestTemplate请求Could not extract response: no suitable HttpMessageConverter found for response type..

Posted 2021-05-18 OkidoGreen

使用 Spring Boot 写项目，需要用到微信接口获取用户信息。

在 Jessey 和 Spring RestTemplate 两个 Rest 客户端中，想到尽量不引入更多的东西，然后就选择了 Spring RestTemplate 作为 网络请求的 Client，然后就被微信接口摆了一道，然后踩了一个 RestTemplate 的坑。

二、第一个坑：被微信摆了一道
报错信息是：

org.springframework.web.client.RestClientException: Could not extract response: no suitable HttpMessageConverter found for response type [class com.solar.app.model.weixin.WxBaseUserInfo] and content type [text/plain]
之所以被微信摆了一道，是因为微信接口文档虽说返回的是 Json 数据，但是同时返回的 Header 里面的 Content-Type 值确是 text/plain 的！！

最终结果就是导致 RestTemplate 把数据从 HttpResponse 转换成 Object 的时候，找不到合适的 HttpMessageConverter 来转换！

我使用 RestTemplate 时配置 Bean 时使用默认的构造函数：

@Bean
RestTemplate restTemplate(){
    return new RestTemplate();
}

继续看 RestTemplate() 默认构造函数都干了啥：

/**
 * Create a new instance of the {@link RestTemplate} using default settings.
 * Default {@link HttpMessageConverter}s are initialized.
 */
public RestTemplate() {
    this.messageConverters.add(new ByteArrayHttpMessageConverter());
    this.messageConverters.add(new StringHttpMessageConverter());
    this.messageConverters.add(new ResourceHttpMessageConverter());
    this.messageConverters.add(new SourceHttpMessageConverter<Source>());
    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());// tag1
    }
    else if (gsonPresent) {
        this.messageConverters.add(new GsonHttpMessageConverter());
    }
}

可以看到，RestTemplate() 默认构造函数设置了一系列 HttpMessageConverter。

我的项目里引入了 com.fasterxml.jackson，所以 RestTemplate() 会构造一个 MappingJackson2HttpMessageConverter 加到它的 messageConverters 中，即上面的代码：【tag1】

继续看 MappingJackson2HttpMessageConverter() 默认构造函数：

/**
 * Construct a new {@link MappingJackson2HttpMessageConverter} using default configuration
 * provided by {@link Jackson2ObjectMapperBuilder}.
 */
public MappingJackson2HttpMessageConverter() {
    this(Jackson2ObjectMapperBuilder.json().build());
}

/**
 * Construct a new {@link MappingJackson2HttpMessageConverter} with a custom {@link ObjectMapper}.
 * You can use {@link Jackson2ObjectMapperBuilder} to build it easily.
 * @see Jackson2ObjectMapperBuilder#json()
 */
public MappingJackson2HttpMessageConverter(ObjectMapper objectMapper) {
    super(objectMapper, MediaType.APPLICATION_JSON, new MediaType("application", "*+json"));
}

可以看到，默认构造的 MappingJackson2HttpMessageConverter 中的 supportedMediaTypes 只支持：application/json 的 MediaType。

再看 RestTemplate 请求的流程，会执行到这里：

/**
 * Execute the given method on the provided URI.
 * <p>The {@link ClientHttpRequest} is processed using the {@link RequestCallback};
 * the response with the {@link ResponseExtractor}.
 * @param url the fully-expanded URL to connect to
 * @param method the HTTP method to execute (GET, POST, etc.)
 * @param requestCallback object that prepares the request (can be {@code null})
 * @param responseExtractor object that extracts the return value from the response (can be {@code null})
 * @return an arbitrary object, as returned by the {@link ResponseExtractor}
 */
protected <T> T doExecute(URI url, HttpMethod method, RequestCallback requestCallback,
        ResponseExtractor<T> responseExtractor) throws RestClientException {

    Assert.notNull(url, "'url' must not be null");
    Assert.notNull(method, "'method' must not be null");
    ClientHttpResponse response = null;
    try {
        ClientHttpRequest request = createRequest(url, method);
        if (requestCallback != null) {
            requestCallback.doWithRequest(request);
        }
        response = request.execute();
        handleResponse(url, method, response);
        if (responseExtractor != null) {
            return responseExtractor.extractData(response);// tag2
        }
        else {
            return null;
        }
    }
    catch (IOException ex) {
        String resource = url.toString();
        String query = url.getRawQuery();
        resource = (query != null ? resource.substring(0, resource.indexOf(query) - 1) : resource);
        throw new ResourceAccessException("I/O error on " + method.name() +
                " request for \\"" + resource + "\\": " + ex.getMessage(), ex);
    }
    finally {
        if (response != null) {
            response.close();
        }
    }
}

从 HttpResponse 中获取数据实际是执行 【tag2】。这个操作由 HttpMessageConverterExtractor 类来完成：

@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);// tag3, 微信返回的是 text/plain

    for (HttpMessageConverter<?> messageConverter : this.messageConverters) {
        if (messageConverter instanceof GenericHttpMessageConverter) {
            GenericHttpMessageConverter<?> genericMessageConverter = (GenericHttpMessageConverter<?>) messageConverter;
            if (genericMessageConverter.canRead(this.responseType, null, contentType)) {// tag4
                if (logger.isDebugEnabled()) {
                    logger.debug("Reading [" + this.responseType + "] as \\"" +
                            contentType + "\\" using [" + messageConverter + "]");
                }
                return (T) genericMessageConverter.read(this.responseType, null, responseWrapper);
            }
        }
        if (this.responseClass != null) {
            if (messageConverter.canRead(this.responseClass, contentType)) {
                if (logger.isDebugEnabled()) {
                    logger.debug("Reading [" + this.responseClass.getName() + "] as \\"" +
                            contentType + "\\" using [" + messageConverter + "]");
                }
                return (T) messageConverter.read((Class) this.responseClass, responseWrapper);
            }
        }
    }

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

【tag4】处的代码用于判断 MappingJackson2HttpMessageConverter 是否支持 【tag3】 类型的 MediaType。

AbstractJackson2HttpMessageConverter：

@Override
public boolean canRead(Type type, Class<?> contextClass, MediaType mediaType) {
    if (!canRead(mediaType)) {// tag5
        return false;
    }
    JavaType javaType = getJavaType(type, contextClass);
    if (!logger.isWarnEnabled()) {
        return this.objectMapper.canDeserialize(javaType);
    }
    AtomicReference<Throwable> causeRef = new AtomicReference<Throwable>();
    if (this.objectMapper.canDeserialize(javaType, causeRef)) {
        return true;
    }
    logWarningIfNecessary(javaType, causeRef.get());
    return false;
}

AbstractHttpMessageConverter：

/**
 * Returns {@code true} if any of the {@linkplain #setSupportedMediaTypes(List)
 * supported} media types {@link MediaType#includes(MediaType) include} the
 * given media type.
 * @param mediaType the media type to read, can be {@code null} if not specified.
 * Typically the value of a {@code Content-Type} header.
 * @return {@code true} if the supported media types include the media type,
 * or if the media type is {@code null}
 */
protected boolean canRead(MediaType mediaType) {
    if (mediaType == null) {
        return true;
    }
    for (MediaType supportedMediaType : getSupportedMediaTypes()) {
        if (supportedMediaType.includes(mediaType)) {
            return true;
        }
    }
    return false;
}

一路追踪下来，可以确定，只要让 MappingJackson2HttpMessageConverter 能处理头部 Content-Type 为 text/plain 类型的 Json 返回值的话，我们就能让其帮我们把 Json 反序列化成我们要的对象。

我们继承 MappingJackson2HttpMessageConverter 并在构造过程中设置其支持的 MediaType 类型即可：

public class WxMappingJackson2HttpMessageConverter extends MappingJackson2HttpMessageConverter {
    public WxMappingJackson2HttpMessageConverter(){
        List<MediaType> mediaTypes = new ArrayList<>();
        mediaTypes.add(MediaType.TEXT_PLAIN);
        setSupportedMediaTypes(mediaTypes);// tag6
    }
}

【tag6】的代码，会覆盖其默认的 MediaType 设置。

然后把这个 WxMappingJackson2HttpMessageConverter 追加到 RestTemplate 的 messageConverters 消息转换链中去：

@Bean
RestTemplate restTemplate(){
    RestTemplate restTemplate = new RestTemplate();
    restTemplate.getMessageConverters().add(new WxMappingJackson2HttpMessageConverter());
    return restTemplate;
}

我既不推荐把 WxMappingJackson2HttpMessageConverter 实例当作构造 RestTemplate 时的参数来构造 RestTemplate，也不推荐 使用新的 WxMappingJackson2HttpMessageConverter 替换 RestTemplate 默认构造中创建的 MappingJackson2HttpMessageConverter 实例，因为这两种方式都会导致 Content-Type 为 application/json 的 Json 响应没有转换器来反序列化，所以最佳的方式还是“追加”。