专栏首页后端沉思录Dubbo SPI实现原理

Dubbo SPI实现原理

Dubbo 并未使用 Java 原生的 SPI 机制,而是对其进行了增强,使其能够更好的满足需求,在 Dubbo 中,SPI 是一个非常重要的模块。基于 SPI,我们可以很容易的对 Dubbo 进行拓展. 本篇文章通过示例说明,先 download 代码,然后在 demo-dubbo--》dubbo-demo-api--》dubbo-demo-api-provider 下新建类:

@SPI("robot")
public interface Robot {
    @Adaptive
    void sayHello();
}

public class Bumblebee implements Robot{
    @Override
    public void sayHello() {
        System.out.println("Hello, I am Bumblebee.");
    }
}

public class OptimusPrime implements Robot{
    @Override
    public void sayHello() {
        System.out.println("Hello, I am Optimus Prime.");
    }
}

然后在 resources 文件夹下创建 META-INF/dubbo/internal ,再在该文件夹下创建文件 org.apache.dubbo.demo.provider.Robot,并写入:

optimusPrime = org.apache.dubbo.demo.provider.OptimusPrime
bumblebee = org.apache.dubbo.demo.provider.Bumblebee

测试方法如下:

public class Test {
    public static void main(String[] args) throws Exception {
        ExtensionLoader<Robot> extensionLoader =
                ExtensionLoader.getExtensionLoader(Robot.class);
        Robot optimusPrime = extensionLoader.getExtension("optimusPrime");
        optimusPrime.sayHello();
        Robot bumblebee = extensionLoader.getExtension("bumblebee");
        bumblebee.sayHello();
    }
}

ExtensionLoader#getExtensionLoader 实现原理

然后运行该 Test 类的main方法,从断点进入 org.apache.dubbo.common.extension.ExtensionLoader#getExtensionLoader 方法,用于获取 ExtensionLoader 实例:

public static <T> ExtensionLoader<T> getExtensionLoader(Class<T> type) {
        // type为空抛出异常
        if (type == null) {
            throw new IllegalArgumentException("Extension type == null");
        }
        // 不是接口抛出异常
        if (!type.isInterface()) {
            throw new IllegalArgumentException("Extension type (" + type + ") is not an interface!");
        }
        // 若没有SPI注解抛出异常
        if (!withExtensionAnnotation(type)) {
            throw new IllegalArgumentException("Extension type (" + type +
                    ") is not an extension, because it is NOT annotated with @" + SPI.class.getSimpleName() + "!");
        }
        // 从缓存EXTENSION_LOADERS获取
        ExtensionLoader<T> loader = (ExtensionLoader<T>) EXTENSION_LOADERS.get(type);
        if (loader == null) {
            EXTENSION_LOADERS.putIfAbsent(type, new ExtensionLoader<T>(type));
            loader = (ExtensionLoader<T>) EXTENSION_LOADERS.get(type);
        }
        return loader;
    }

首选会检查type是否为空、type是否为接口、type是否有SPI注解(Robot必须有SPI注解,否则会报错),再从缓存中获取 ExtensionLoader,如果为空,从新创建ExtensionLoader实例,断点进入 ExtensionLoader 构造函数:

private ExtensionLoader(Class<?> type) {
        this.type = type;
        objectFactory = (type == ExtensionFactory.class ? null : ExtensionLoader.getExtensionLoader(ExtensionFactory.class).getAdaptiveExtension());
    }

如果type是 ExtensionFactory.class 时,objectFactory初始化为null,否则执行 ExtensionLoader.getExtensionLoader(ExtensionFactory.class).getAdaptiveExtension() ,断点继续进入,还是会进入 org.apache.dubbo.common.extension.ExtensionLoader#getExtensionLoader 方法,这里就不做分析,由于此时type为 ExtensionFactory.class ,所以objectFactory为null,这里分析 org.apache.dubbo.common.extension.ExtensionLoader#getAdaptiveExtension ,断点进入:

// cachedAdaptiveInstance
        Object instance = cachedAdaptiveInstance.get();
        if (instance == null) {
            // createAdaptiveInstanceError不为null则报错
            if (createAdaptiveInstanceError != null) {
                throw new IllegalStateException("Failed to create adaptive instance: " +
                        createAdaptiveInstanceError.toString(),
                        createAdaptiveInstanceError);
            }
            // dubbo里常出现的双重检查
            synchronized (cachedAdaptiveInstance) {
                instance = cachedAdaptiveInstance.get();
                if (instance == null) {
                    try {
                        // instance为空,则创建
                        instance = createAdaptiveExtension();
                        cachedAdaptiveInstance.set(instance);
                    } catch (Throwable t) {
                        createAdaptiveInstanceError = t;
                        throw new IllegalStateException("Failed to create adaptive instance: " + t.toString(), t);
                    }
                }
            }
        }

        return (T) instance;

断点进入 org.apache.dubbo.common.extension.ExtensionLoader#createAdaptiveExtension:

private T createAdaptiveExtension() {
        try {
            return injectExtension((T) getAdaptiveExtensionClass().newInstance());
        } catch (Exception e) {
            throw new IllegalStateException("Can't create adaptive extension " + type + ", cause: " + e.getMessage(), e);
        }
    }

injectExtension 负责扩展点的依赖注入, getAdaptiveExtensionClass 方法为了获取自适应扩展类,断点进入 org.apache.dubbo.common.extension.ExtensionLoader#getAdaptiveExtensionClass :

private Class<?> getAdaptiveExtensionClass() {
        getExtensionClasses();
        if (cachedAdaptiveClass != null) {
            return cachedAdaptiveClass;
        }
        return cachedAdaptiveClass = createAdaptiveExtensionClass();
    }

无疑继续进入方法 getExtensionClasses ,获取扩展类:

private Map<String, Class<?>> getExtensionClasses() {
        Map<String, Class<?>> classes = cachedClasses.get();
        if (classes == null) {
            synchronized (cachedClasses) {
                classes = cachedClasses.get();
                if (classes == null) {
                    classes = loadExtensionClasses();
                    cachedClasses.set(classes);
                }
            }
        }
        return classes;
    }

这里也没什么好说的,首先从缓存里获取,缓存里没有则调用 loadExtensionClasses 加载,断点进入:

private Map<String, Class<?>> loadExtensionClasses() {
        cacheDefaultExtensionName();

        Map<String, Class<?>> extensionClasses = new HashMap<>();
        loadDirectory(extensionClasses, DUBBO_INTERNAL_DIRECTORY, type.getName());
        loadDirectory(extensionClasses, DUBBO_INTERNAL_DIRECTORY, type.getName().replace("org.apache", "com.alibaba"));
        loadDirectory(extensionClasses, DUBBO_DIRECTORY, type.getName());
        loadDirectory(extensionClasses, DUBBO_DIRECTORY, type.getName().replace("org.apache", "com.alibaba"));
        loadDirectory(extensionClasses, SERVICES_DIRECTORY, type.getName());
        loadDirectory(extensionClasses, SERVICES_DIRECTORY, type.getName().replace("org.apache", "com.alibaba"));
        return extensionClasses;
    }

会从目录

META-INF/dubbo/internal/
META-INF/dubbo/
META-INF/services/

中根据type名称加载文件,此时type.getName为 org.apache.dubbo.common.extension.ExtensionFactory ,继续进入 loadDirectory :

private void loadDirectory(Map<String, Class<?>> extensionClasses, String dir, String type) {
        String fileName = dir + type;
        try {
            Enumeration<java.net.URL> urls;
            ClassLoader classLoader = findClassLoader();
            if (classLoader != null) {
                urls = classLoader.getResources(fileName);
            } else {
                urls = ClassLoader.getSystemResources(fileName);
            }
            if (urls != null) {
                while (urls.hasMoreElements()) {
                    java.net.URL resourceURL = urls.nextElement();
                    loadResource(extensionClasses, classLoader, resourceURL);
                }
            }
        } catch (Throwable t) {
            logger.error("Exception occurred when loading extension class (interface: " +
                    type + ", description file: " + fileName + ").", t);
        }
    }

和上面所说一样,根据名称获取资源文件,然后调用 loadResource 加载,进入该方法:

private void loadResource(Map<String, Class<?>> extensionClasses, ClassLoader classLoader, java.net.URL resourceURL) {
        try {
            try (BufferedReader reader = new BufferedReader(new InputStreamReader(resourceURL.openStream(), StandardCharsets.UTF_8))) {
                String line;
                while ((line = reader.readLine()) != null) {
                    final int ci = line.indexOf('#');
                    if (ci >= 0) {
                        line = line.substring(0, ci);
                    }
                    line = line.trim();
                    if (line.length() > 0) {
                        try {
                            String name = null;
                            int i = line.indexOf('=');
                            if (i > 0) {
                                name = line.substring(0, i).trim();
                                line = line.substring(i + 1).trim();
                            }
                            if (line.length() > 0) {
                                loadClass(extensionClasses, resourceURL, Class.forName(line, true, classLoader), name);
                            }
                        } catch (Throwable t) {
                            IllegalStateException e = new IllegalStateException("Failed to load extension class (interface: " + type + ", class line: " + line + ") in " + resourceURL + ", cause: " + t.getMessage(), t);
                            exceptions.put(line, e);
                        }
                    }
                }
            }
        } catch (Throwable t) {
            logger.error("Exception occurred when loading extension class (interface: " +
                    type + ", class file: " + resourceURL + ") in " + resourceURL, t);
        }
    }

上面代码获取到文件流后,然后读取解析得到key、value值,继续调用 loadClass 方法,断点进入:

private void loadClass(Map<String, Class<?>> extensionClasses, java.net.URL resourceURL, Class<?> clazz, String name) throws NoSuchMethodException {
        if (!type.isAssignableFrom(clazz)) {
            throw new IllegalStateException("Error occurred when loading extension class (interface: " +
                    type + ", class line: " + clazz.getName() + "), class "
                    + clazz.getName() + " is not subtype of interface.");
        }
        // 判断类上是否带Adaptive注解,如果带的话,把clazz赋值给cachedAdaptiveClass
        if (clazz.isAnnotationPresent(Adaptive.class)) {
            cacheAdaptiveClass(clazz);
        } else if (isWrapperClass(clazz)) {
            // cache wrapper class
            cacheWrapperClass(clazz);
        } else {
            clazz.getConstructor();
            if (StringUtils.isEmpty(name)) {
                name = findAnnotationName(clazz);
                if (name.length() == 0) {
                    throw new IllegalStateException("No such extension name for the class " + clazz.getName() + " in the config " + resourceURL);
                }
            }

            String[] names = NAME_SEPARATOR.split(name);
            if (ArrayUtils.isNotEmpty(names)) {
                // 如果类带有Activate注解,则缓存
                cacheActivateClass(clazz, names[0]);
                for (String n : names) {
                    cacheName(clazz, n);
                    // 保存扩展类
                    saveInExtensionClass(extensionClasses, clazz, n);
                }
            }
        }
    }

上面注意 org.apache.dubbo.common.extension.ExtensionLoader#isWrapperClass 方法,判断是否为包装类,是的话就缓存,判断依据就是扩展接口为该类的构造器参数,例如 Protocol 的两个子类 org.apache.dubbo.rpc.protocol.ProtocolFilterWrapperorg.apache.dubbo.rpc.protocol.ProtocolListenerWrapper ,它们的构造函数都有以 Protocol 为参数,所以为包装类,在 org.apache.dubbo.common.extension.ExtensionLoader#createExtension 中会判断是否有包装类,有的的话会循环注入,例如在获取对象 org.apache.dubbo.rpc.Protocol 时,通过自适应扩展获取到的对象会被包装类包裹,例如:

org.apache.dubbo.rpc.protocol.ProtocolFilterWrapper
     org.apache.dubbo.rpc.protocol.ProtocolListenerWrapper
          org.apache.dubbo.registry.integration.RegistryProtocol

这样的好处就是在调用前可以添加额外的处理逻辑.

到这里扩展类已经加载完了,并且保存,回到 org.apache.dubbo.common.extension.ExtensionLoader#getAdaptiveExtensionClass方法中,调用完 getExtensionClasses 后,会判断 cachedAdaptiveClass 是否为空,如果不为空直接返回,也就是如果类上有 Adaptive注解,此时会直接返回,注意结合 org.apache.dubbo.common.extension.ExtensionLoader#loadClass 方法中的判断分析,否则继续执行调用 createAdaptiveExtensionClass ,但是此时 cachedAdaptiveClass 是不为空的,所以直接返回.为什么不为空呢,由于 ExtensionFactory 的实现类 org.apache.dubbo.common.extension.factory.AdaptiveExtensionFactory 加上了 Adaptive 注解,在 org.apache.dubbo.common.extension.ExtensionLoader#loadClass方法中,初始化了 cachedAdaptiveClass ,dubbo中的类上被 Adaptive 修饰的非常少,仅有两个: AdaptiveCompilerAdaptiveExtensionFactory.

回到 org.apache.dubbo.common.extension.ExtensionLoader#createAdaptiveExtension 方法中, getAdaptiveExtensionClass 获取到扩展类后,通过反射创建对象,然后再调用 injectExtension ,断点进入:

private T injectExtension(T instance) {
        // 若objectFactory为null,直接返回instance
        if (objectFactory == null) {
            return instance;
        }
        try {
            for (Method method : instance.getClass().getMethods()) {
                if (!isSetter(method)) {
                    continue;
                }
                /**
                 * Check {@link DisableInject} to see if we need auto injection for this property
                 */
                if (method.getAnnotation(DisableInject.class) != null) {
                    continue;
                }
                Class<?> pt = method.getParameterTypes()[0];
                if (ReflectUtils.isPrimitives(pt)) {
                    continue;
                }
                try {
                    String property = getSetterProperty(method);
                    Object object = objectFactory.getExtension(pt, property);
                    if (object != null) {
                        method.invoke(instance, object);
                    }
                } catch (Exception e) {
                    logger.error("Failed to inject via method " + method.getName()
                            + " of interface " + type.getName() + ": " + e.getMessage(), e);
                }
            }
        } catch (Exception e) {
            logger.error(e.getMessage(), e);
        }
        return instance;
    }

首先判断 objectFactory 是否为空,为空则直接返回,前面分析的 ExtensionLoader 构造方法的时若扩展点类型是 ExtensionFactory ,则 objectFactory 为null,否则为 objectFactory 的自适应扩展,如果 objectFactory 不为空,则遍历所有的 setter 方法,如果方法上有 @DisableInject 则直接跳过,否则通过 objectFactory 获取对应的属性,不为空则调用 setter 方法,如果为 SPI 扩展点注入属性值. 回到 org.apache.dubbo.common.extension.ExtensionLoader#getExtensionLoader 方法中, ExtensionLoader 实例已创建完成,并放入 EXTENSION_LOADERS 中,key是 interfaceorg.apache.dubbo.demo.provider.Robot.class ,value是 ExtensionLoader 对象.

ExtensionLoader#getExtension实现

回到 org.apache.dubbo.demo.provider.Test#main 中,即 ExtensionLoader 对象已经创建成功,下一步获取扩展对象:

Robot optimusPrime = extensionLoader.getExtension("optimusPrime");

断点进入 org.apache.dubbo.common.extension.ExtensionLoader#getExtension 中:

public T getExtension(String name) {
        // 若扩展名称为空,则抛出异常
        if (StringUtils.isEmpty(name)) {
            throw new IllegalArgumentException("Extension name == null");
        }
        // 若扩展名为true,返回默认扩展对象
        if ("true".equals(name)) {
            return getDefaultExtension();
        }
        // 获取Holder对象,如果从缓存获取为空,则重新创建
        final Holder<Object> holder = getOrCreateHolder(name);
        Object instance = holder.get();
        if (instance == null) {
            synchronized (holder) {
                instance = holder.get();
                if (instance == null) {
                    instance = createExtension(name);
                    holder.set(instance);
                }
            }
        }
        return (T) instance;
    }

上面代码主要看 org.apache.dubbo.common.extension.ExtensionLoader#createExtension 方法,断点进入:

private T createExtension(String name) {
        Class<?> clazz = getExtensionClasses().get(name);
        if (clazz == null) {
            throw findException(name);
        }
        try {
            T instance = (T) EXTENSION_INSTANCES.get(clazz);
            if (instance == null) {
                EXTENSION_INSTANCES.putIfAbsent(clazz, clazz.newInstance());
                instance = (T) EXTENSION_INSTANCES.get(clazz);
            }
            // 依赖注入属性
            injectExtension(instance);
            Set<Class<?>> wrapperClasses = cachedWrapperClasses;
            // Wrapper包装
            if (CollectionUtils.isNotEmpty(wrapperClasses)) {
                for (Class<?> wrapperClass : wrapperClasses) {
                    instance = injectExtension((T) wrapperClass.getConstructor(type).newInstance(instance));
                }
            }
            return instance;
        } catch (Throwable t) {
            throw new IllegalStateException("Extension instance (name: " + name + ", class: " +
                    type + ") couldn't be instantiated: " + t.getMessage(), t);
        }
    }

明显继续进入 org.apache.dubbo.common.extension.ExtensionLoader#getExtensionClasses 方法:

private Map<String, Class<?>> getExtensionClasses() {
        Map<String, Class<?>> classes = cachedClasses.get();
        if (classes == null) {
            synchronized (cachedClasses) {
                classes = cachedClasses.get();
                if (classes == null) {
                    classes = loadExtensionClasses();
                    cachedClasses.set(classes);
                }
            }
        }
        return classes;
    }

双重检查判断,继续断点进入 org.apache.dubbo.common.extension.ExtensionLoader#loadExtensionClasses :

private Map<String, Class<?>> loadExtensionClasses() {
        cacheDefaultExtensionName();

        Map<String, Class<?>> extensionClasses = new HashMap<>();
        loadDirectory(extensionClasses, DUBBO_INTERNAL_DIRECTORY, type.getName());
        loadDirectory(extensionClasses, DUBBO_INTERNAL_DIRECTORY, type.getName().replace("org.apache", "com.alibaba"));
        loadDirectory(extensionClasses, DUBBO_DIRECTORY, type.getName());
        loadDirectory(extensionClasses, DUBBO_DIRECTORY, type.getName().replace("org.apache", "com.alibaba"));
        loadDirectory(extensionClasses, SERVICES_DIRECTORY, type.getName());
        loadDirectory(extensionClasses, SERVICES_DIRECTORY, type.getName().replace("org.apache", "com.alibaba"));
        return extensionClasses;
    }

这段代码是不是似曾相识,不错,就是上面创建 ExtensionLoader 时加载自适应扩展类时调用的方法,思路和上面的一样,就不再次分析了,解析完得到 的 Map 里存放的是:

optimusPrime -> {Class@1527} "class org.apache.dubbo.demo.provider.OptimusPrime"
bumblebee -> {Class@1569} "class org.apache.dubbo.demo.provider.Bumblebee"

回到 org.apache.dubbo.common.extension.ExtensionLoader#createExtension方法内,此时根据 optimusPrime 就可以从 Map 中获取到 org.apache.dubbo.demo.provider.OptimusPrime.class,判断 EXTENSION_INSTANCES 中是否有缓存,没有的话会根据反射创建该类对象,并放入到 EXTENSION_INSTANCES ,然后调用方法 org.apache.dubbo.common.extension.ExtensionLoader#injectExtension,依赖注入属性,这个不作分析,与上文一致.至此对象 OptimusPrime 已生成了,然后返回到 Test 类测试方法中,可调用具体的方法了. 通过 ExtensionLoader#getExtension获取扩展点的实现,并不能体现自适应特性,和java SPI 并没有多大的差别,下面介绍 ExtensionLoader#getAdaptiveExtension 的实现.

ExtensionLoader#getAdaptiveExtension的实现

下面的分析需要注册中心 zookeeper ,关于安装和配置就不多分析.首先在 `org.apache.dubbo.config.ServiceConfig#protocol 处打上断点:

private static final Protocol protocol = ExtensionLoader.getExtensionLoader(Protocol.class).getAdaptiveExtension();

在加载类 ServiceConfig 时,会先初始化静态变量 protocol , 而此时会获取自适应扩展类.到模块 demo-dubbo--》dubbo-demo-api--》dubbo-demo-api-provider 中运行 Application类的main方法:

public class Application {
    public static void main(String[] args) throws Exception {
        ServiceConfig<DemoServiceImpl> service = new ServiceConfig<>();
        service.setApplication(new ApplicationConfig("dubbo-demo-api-provider"));
        service.setRegistry(new RegistryConfig("zookeeper://127.0.0.1:2181"));
        service.setInterface(DemoService.class);
        service.setRef(new DemoServiceImpl());
        service.export();
        System.in.read();
    }
}

从断点进入,关于 ExtensionLoader#getExtensionLoader 的实现上文已经分析,这里不多作分析,断点直接进入 org.apache.dubbo.common.extension.ExtensionLoader#getAdaptiveExtensionClass 方法内:

private Class<?> getAdaptiveExtensionClass() {
        getExtensionClasses();
        if (cachedAdaptiveClass != null) {
            return cachedAdaptiveClass;
        }
        return cachedAdaptiveClass = createAdaptiveExtensionClass();
    }

由上文分析可知,由于 Protocol 的实现类上都没有注解 Adaptive ,所以 cachedAdaptiveClass 为空,此时会调用方法 org.apache.dubbo.common.extension.ExtensionLoader#createAdaptiveExtensionClass ,断点进入:

private Class<?> createAdaptiveExtensionClass() {
        // 获取扩展类字符串
        String code = new AdaptiveClassCodeGenerator(type, cachedDefaultName).generate();
        ClassLoader classLoader = findClassLoader();
        org.apache.dubbo.common.compiler.Compiler compiler = ExtensionLoader.getExtensionLoader(org.apache.dubbo.common.compiler.Compiler.class).getAdaptiveExtension();
        return compiler.compile(code, classLoader);
    }

断点进入 org.apache.dubbo.common.extension.AdaptiveClassCodeGenerator#generate :

public String generate() {
        // no need to generate adaptive class since there's no adaptive method found.
        // type中至少一个方法被注解Adaptive修饰
        if (!hasAdaptiveMethod()) {
            throw new IllegalStateException("No adaptive method exist on extension " + type.getName() + ", refuse to create the adaptive class!");
        }

        StringBuilder code = new StringBuilder();
        // 拼接包名
        code.append(generatePackageInfo());
        // 拼接依赖包
        code.append(generateImports());
        // 拼接类名
        code.append(generateClassDeclaration());

        Method[] methods = type.getMethods();
        // 拼接方法
        for (Method method : methods) {
            code.append(generateMethod(method));
        }
        code.append("}");

        if (logger.isDebugEnabled()) {
            logger.debug(code.toString());
        }
        return code.toString();
    }

首先会校验 扩展类的接口中至少有一个方法被 Adaptive 注解修饰,具体代码不作分析了,看生成的类字符串信息:

package org.apache.dubbo.rpc;
    import org.apache.dubbo.common.extension.ExtensionLoader;
    public class Protocol$Adaptive implements org.apache.dubbo.rpc.Protocol {
        public void destroy() {
            throw new UnsupportedOperationException("The method public abstract void org.apache.dubbo.rpc.Protocol.destroy() of interface org.apache.dubbo.rpc.Protocol is not adaptive method!");
        }

        public int getDefaultPort() {
            throw new UnsupportedOperationException("The method public abstract int org.apache.dubbo.rpc.Protocol.getDefaultPort() of interface org.apache.dubbo.rpc.Protocol is not adaptive method!");
        }

        public org.apache.dubbo.rpc.Invoker refer(java.lang.Class arg0, org.apache.dubbo.common.URL arg1) throws org.apache.dubbo.rpc.RpcException {
            if (arg1 == null) {
                throw new IllegalArgumentException("url == null");
            }
            org.apache.dubbo.common.URL url = arg1;
            String extName = (url.getProtocol() == null ? "dubbo" : url.getProtocol());
            if (extName == null) {
                throw new IllegalStateException("Failed to get extension (org.apache.dubbo.rpc.Protocol) name from url (" + url.toString() + ") use keys([protocol])");
            }
            org.apache.dubbo.rpc.Protocol extension = (org.apache.dubbo.rpc.Protocol) ExtensionLoader.getExtensionLoader(org.apache.dubbo.rpc.Protocol.class).getExtension(extName);
            return extension.refer(arg0, arg1);
        }

        public org.apache.dubbo.rpc.Exporter export(org.apache.dubbo.rpc.Invoker arg0) throws org.apache.dubbo.rpc.RpcException {
            if (arg0 == null) {
                throw new IllegalArgumentException("org.apache.dubbo.rpc.Invoker argument == null");
            }
            if (arg0.getUrl() == null) {
                throw new IllegalArgumentException("org.apache.dubbo.rpc.Invoker argument getUrl() == null");
            }
            org.apache.dubbo.common.URL url = arg0.getUrl();
            String extName = (url.getProtocol() == null ? "dubbo" : url.getProtocol());
            if (extName == null) {
                throw new IllegalStateException("Failed to get extension (org.apache.dubbo.rpc.Protocol) name from url (" + url.toString() + ") use keys([protocol])");
            }
            org.apache.dubbo.rpc.Protocol extension = (org.apache.dubbo.rpc.Protocol) ExtensionLoader.getExtensionLoader(org.apache.dubbo.rpc.Protocol.class).getExtension(extName);
            return extension.export(arg0);
        }
    }

上面生成了类Protocol的子类 Protocol$Adaptive ,仔细观察发发现, exportrefer 方法都被注解 Adaptive 修饰,生成的 Protocol$Adaptive 中这两个方法都有具体的实现,而 destroygetDefaultPort 没有被注解修饰,生成的类中直接抛出 UnsupportedOperationException 异常. 回到 org.apache.dubbo.common.extension.ExtensionLoader#createAdaptiveExtensionClass 方法中, 由于 org.apache.dubbo.common.compiler.Compiler 的子类 org.apache.dubbo.common.compiler.support.AdaptiveCompiler 被 注解 Adaptive 修饰,所以变量 compiler 的值是 AdaptiveCompiler 实例,断点进入 org.apache.dubbo.common.compiler.support.AdaptiveCompiler#compile:

public Class<?> compile(String code, ClassLoader classLoader) {
        Compiler compiler;
        // 获取ExtensionLoader
        ExtensionLoader<Compiler> loader = ExtensionLoader.getExtensionLoader(Compiler.class);
        String name = DEFAULT_COMPILER; // copy reference
        if (name != null && name.length() > 0) {
            compiler = loader.getExtension(name);
        } else {
            compiler = loader.getDefaultExtension();
        }
        return compiler.compile(code, classLoader);
    }

上面首先或获取 ExtensionLoader, 根据 name 判断去加载具体的扩展实现,这里由于name为空,所以会调用 org.apache.dubbo.common.extension.ExtensionLoader#getDefaultExtension,断点进入:

public T getDefaultExtension() {
        // 获取扩展类
        // 上一步已经获取到并且放入缓存,如下:
        // jdk -> {Class@1244} "class org.apache.dubbo.common.compiler.support.JdkCompiler"
        // javassist -> {Class@1245} "class org.apache.dubbo.common.compiler.support.JavassistCompiler"
        //
        getExtensionClasses();
        if (StringUtils.isBlank(cachedDefaultName) || "true".equals(cachedDefaultName)) {
            return null;
        }
        // 默认获取 javassist
        return getExtension(cachedDefaultName);
    }

org.apache.dubbo.common.extension.ExtensionLoader#getExtension 的具体实现不作分析,上文已经分析过,经过该方法后,可以得到 org.apache.dubbo.common.compiler.support.JavassistCompiler 实例. 然后回到 org.apache.dubbo.common.compiler.support.AdaptiveCompiler#compile方法内,执行 org.apache.dubbo.common.compiler.support.AbstractCompiler#compile 方法,该方法是父类方法,做一些检验处理,然后再调用模版方法 doCompile 由子类AdaptiveCompiler#doCompile 去处理,底层由 Javassist 生成类 Protocol$Adaptive,到此自适应扩展类的生成已分析完毕.但是生成完如何使用呢?

自适应扩展按需加载

回到 org.apache.dubbo.config.ServiceConfig中,在此处打上断点:

Exporter<?> exporter = protocol.export(wrapperInvoker);

wrapperInvoker对象包装着invoker,具体可以看截图:

我们主要关注URL对象里的 string :

registry://127.0.0.1:2181/org.apache.dubbo.registry.RegistryService?application=dubbo-demo-api-provider&dubbo=2.0.2&export=dubbo%3A%2F%2F10.200.190.156%3A20880%2Forg.apache.dubbo.demo.DemoService%3Fanyhost%3Dtrue%26application%3Ddubbo-demo-api-provider%26bind.ip%3D10.200.190.156%26bind.port%3D20880%26deprecated%3Dfalse%26dubbo%3D2.0.2%26dynamic%3Dtrue%26generic%3Dfalse%26interface%3Dorg.apache.dubbo.demo.DemoService%26methods%3DsayHello%26pid%3D21663%26release%3D%26side%3Dprovider%26timestamp%3D1568697592181&pid=21663&registry=zookeeper&timestamp=1568697592172

很明显使用的协议: registry,回到生成的自适应扩展类 Protocol$Adaptive 中的 export 方法:

public org.apache.dubbo.rpc.Exporter export(org.apache.dubbo.rpc.Invoker arg0) throws org.apache.dubbo.rpc.RpcException {
            if (arg0 == null) {
                throw new IllegalArgumentException("org.apache.dubbo.rpc.Invoker argument == null");
            }
            if (arg0.getUrl() == null) {
                throw new IllegalArgumentException("org.apache.dubbo.rpc.Invoker argument getUrl() == null");
            }
            org.apache.dubbo.common.URL url = arg0.getUrl();
            String extName = (url.getProtocol() == null ? "dubbo" : url.getProtocol());
            if (extName == null) {
                throw new IllegalStateException("Failed to get extension (org.apache.dubbo.rpc.Protocol) name from url (" + url.toString() + ") use keys([protocol])");
            }
            org.apache.dubbo.rpc.Protocol extension = (org.apache.dubbo.rpc.Protocol) ExtensionLoader.getExtensionLoader(org.apache.dubbo.rpc.Protocol.class).getExtension(extName);
            return extension.export(arg0);
        }

首先会判断传入参数是否为空,然后再判断 org.apache.dubbo.common.URL 是否为空, 然后获取协议, 所以这里 extNameregistry,后面就是根据 extName 获取具体的自适应扩展类,很明显 extension 会被 org.apache.dubbo.registry.integration.RegistryProtocol 初始化.

总结

DUBBO SPI解决了什么问题呢,官方文档给出了解释:

  • JDK 标准的 SPI 会一次性实例化扩展点所有实现,如果有扩展实现初始化很耗时,但如果没用上也加载,会很浪费资源
  • 如果扩展点加载失败,连扩展点的名称都拿不到了。比如:JDK 标准的 ScriptEngine,通过 getName() 获取脚本类型的名称,但如果 RubyScriptEngine 因为所依赖的 jruby.jar 不存在,导致 RubyScriptEngine 类加载失败,这个失败原因被吃掉了,和 ruby 对应不起来,当用户执行 ruby 脚本时,会报不支持 ruby,而不是真正失败的原因
  • 增加了对扩展点 IoC 和 AOP 的支持,一个扩展点可以直接 setter 注入其它扩展点

参考文章

  1. http://dubbo.apache.org/zh-cn/docs/sourcecodeguide/dubbo-spi.html
  2. http://dubbo.apache.org/zh-cn/docs/sourcecodeguide/adaptive-extension.html
  3. http://dubbo.apache.org/zh-cn/docs/dev/SPI.html
  4. http://dubbo.apache.org/zh-cn/docs/sourcecodeguide/adaptive-extension.html
  5. https://cxis.me/2017/02/18/Dubbo%E4%B8%ADSPI%E6%89%A9%E5%B1%95%E6%9C%BA%E5%88%B6%E8%AF%A6%E8%A7%A3/
  6. https://kexianjun.github.io/2019/08/19/spi/#more

本文分享自微信公众号 - 后端沉思录(LiosWong),作者:LiosWong

原文出处及转载信息见文内详细说明,如有侵权,请联系 yunjia_community@tencent.com 删除。

原始发表时间:2019-09-23

本文参与腾讯云自媒体分享计划,欢迎正在阅读的你也加入,一起分享。

我来说两句

0 条评论
登录 后参与评论

相关文章

  • dubbo消费者启动报错分析

    通过图可知 org.apache.dubbo.registry.integration.RegistryProtocol#destroy该方法都会在 org.a...

    LiosWong
  • arthas简单使用

    Arthas 是Alibaba开源的Java诊断工具,深受开发者喜爱,项目地址.当你遇到以下类似问题而束手无策时,Arthas可以帮助你解决:

    LiosWong
  • Dubbo服务暴露过程解析

    Dubbo SPI的暴露原理参考https://lioswong.github.io/2019/09/17/Dubbo-SPI%E5%AE%9E%E7%8E%B...

    LiosWong
  • dubbo+zookeeper+springBoot框架整合与dubbo泛型调用演示

    利用maven打dubbo-admin-0.0.1-SNAPSHOT.jar包 8

    授客
  • jib打包docker镜像实战

    codecraft
  • Dubbo 与 Istio 共生

    当下,越来越多的应用迁入了容器集群。“享受着”云原生带来的便利。mesh 也越来越被广大程序员所接受。

    谢正伟
  • Dubbo 与 Istio 共生

    当下,越来越多的应用迁入了容器集群。“享受着”云原生带来的便利。mesh 也越来越被广大程序员所接受。

    谢正伟
  • 腾讯云年末特惠

    而且都是超低价了,就拿第一个4核16G的来说,官网原价是8040/年,5年就是4万多,一口价直接就便宜了3万元,简直美滋滋!

    福利活动清单
  • WAMP环境配置

    经过以上的步骤,一个 WAMP 的开发环境就配置好了,基本上能够满足开发一般应用的需要。

    大江小浪
  • 数学建模竞赛(国赛和美赛)经验分享

    第一次参赛是在大一的暑假参加的国赛,当时和两个同学刚刚组队,我们也没有什么基础,结果可想而知:无奖。 在经历了这一次国赛之后,大一时的两位队友也无心再参加,所以...

    code随笔

扫码关注云+社区

领取腾讯云代金券