JAVA SPI

概述

SPI(Service Provider Interface)，服务提供者接口，它是JDK提供的一种服务发现机制。目前在很多开源项目中使用，如common-logging，JDBC等。

举个例子，我们在项目开发中经常会用到各种序列化，一般情况下，我们会定义好接口，再继承接口实现具体逻辑，如hession。如果一段时间后想替换成protobuffer，使用SPI机制，就可以通过配置文件可以实现动态替换。因此SPI是"基于接口编程＋策略模式＋配置文件"组合实现的动态加载机制。

实现：

1. 定义序列化接口

public interface Serializer {
    /**
     * 序列化
     * @param msg
     * @return
     * @throws IOException
     */
    byte[] serialize(Object msg) throws IOException;

    /**
     * 反序列化
     * @param data
     * @param type
     * @param <T>
     * @return
     * @throws IOException
     */
    <T> T deserialize(byte[] data, Class<T> type) throws IOException;
}

2. 继承Serializer接口，实现fasjson，hessian，kryo和protostuff四种具体序列化方式

public class ProtostuffSerializer implements Serializer {
    private static Map<Class<?>, Schema<?>> cachedSchema = new ConcurrentHashMap<>();
    private static Objenesis objenesis = new ObjenesisStd(true);

    @Override
    public byte[] serialize(Object msg) throws IOException {
        LinkedBuffer buffer = null;
        try {
            buffer = LinkedBuffer.allocate(LinkedBuffer.DEFAULT_BUFFER_SIZE);
            Schema schema = getSchema(msg.getClass());
            return ProtostuffIOUtil.toByteArray(msg, schema, buffer);
        }catch (Exception e){
            throw e;
        }finally {
            if (buffer != null) {
                buffer.clear();
            }
        }
    }

    @Override
    public <T> T deserialize(byte[] data, Class<T> clazz) throws IOException {
        try {
            T message = objenesis.newInstance(clazz);
            Schema schema = getSchema(clazz);
            ProtostuffIOUtil.mergeFrom(data, message, schema);
            return message;
        }catch (Exception e){
            throw e;
        }
    }

    private static <T> Schema getSchema(Class<T> clazz){
        Schema schema = cachedSchema.get(clazz);
        if(schema == null){
            schema = RuntimeSchema.createFrom(clazz);
            cachedSchema.put(clazz, schema);
        }

        return schema;
    }
}

3. 在resources目录下新建META-INF/services目录，文件名是接口全限定名：com.*.serializer.Serializer，文件内容为具体实现类的全限定名：com.*.serializer.protostuff.ProtostuffSerializer

4. Serializer接口的具体实现类就是文件中定义的实现

Serializer serializer = SpiLoderFactory.getInstance(Serializer.class);

public static <T> T getInstance(Class<?> clazz){
    return (T) ServiceLoader.load(clazz).iterator().next();
}

5. ServiceLoader类中定义了文件读取的目录，也就是为什么我们自定义的文件要存在于META-INF/services下

6. 基本原理还是通过反射将对象放到providers这个linkedHashMap中，通过遍历即可访问

private S nextService() {
    if (!hasNextService())
        throw new NoSuchElementException();
    String cn = nextName;
    nextName = null;
    Class<?> c = null;
    try {
        c = Class.forName(cn, false, loader);
    } catch (ClassNotFoundException x) {
        fail(service,
             "Provider " + cn + " not found");
    }
    if (!service.isAssignableFrom(c)) {
        fail(service,
             "Provider " + cn  + " not a subtype");
    }
    try {
        S p = service.cast(c.newInstance());
        providers.put(cn, p);
        return p;
    } catch (Throwable x) {
        fail(service,
             "Provider " + cn + " could not be instantiated",
             x);
    }
    throw new Error();          // This cannot happen
}