Spring 注解组合实现原理

我们在Spring和Spring boot使用过程中会发现，我们时常会用@RestController代替@Controller+@ResponseBody（很可惜，很多人还在混用）通过看@RestController

@Target(ElementType.TYPE)
@Retention(RetentionPolicy.RUNTIME)
@Documented
@Controller
@ResponseBody
public @interface RestController {

	/**
	 * The value may indicate a suggestion for a logical component name,
	 * to be turned into a Spring bean in case of an autodetected component.
	 * @return the suggested component name, if any (or empty String otherwise)
	 * @since 4.0.1
	 */
	String value() default "";

}

明显的，@RestController包含了@Controller+@ResponseBody，这样也可以！？！？！？是的

看下Spring的考量：

大量的注解虽然简化了配置（如@Controller+@ResponseBody），但是一旦多了起来，同样难以清晰，所以，Spring提供了一个专用的ClassReader，可以解析注解上的注解，便于我们通过场景进行组合使用，而对外来说，就只是一个语义明确的注解（@RestController）。

同样的，对于所有使用到注解的地方，均可以这样使用，比如，假设我们的配置类有一个通用的启用逻辑：

• 条件1成立
• 条件2成立
• .......

首先我们想到的是集成Condition，并实现条件判断，这样，n个Condition就开发出来了，接下里使用时就是有几个配置类每个写一坨

@XXCondition
@YYCondition
...
public class ZConfiguration{
    ...
}

@XXCondition
@YYCondition
...
public class WConfiguration{
    ...
}
...

然后，想到了一个虚拟注解，仅用一个来代表

@Target(ElementType.TYPE)
@Retention(RetentionPolicy.RUNTIME)
@Documented
@XXCondition
@YYCOndition
...
public @interface TotalCondition {

	...

}

so，我们的配置类就变成了这个样子

@TotalCondition
public class ZConfiguration{
    ...
}

@TotalCondition
public class WConfiguration{
    ...
}
...

好了，今天就到这儿了

 

附 代码实现：

废话不多说，直接上！！先来了一段比较难看的代码

Spring在初始化bean时其中一步是获取bean的MetadataVisitor，用于将class文件中该类的注解信息、属性等解析出来，今天我们要说的就是注解的解析，以下方法为总的入口

org.springframework.asm.ClassReader#accept(org.springframework.asm.ClassVisitor, org.springframework.asm.Attribute[], int)

......
    // visits the class annotations and type annotations
    if (ANNOTATIONS && anns != 0) {
        for (int i = readUnsignedShort(anns), v = anns + 2; i > 0; --i) {
            // row num 676
            v = readAnnotationValues(v + 2, c, true,
                            classVisitor.visitAnnotation(readUTF8(v, c), true));
        }
    }
......

    /**
     * Reads the values of an annotation and makes the given visitor visit them.
     * 
     * @param v
     *            the start offset in {@link #b b} of the values to be read
     *            (including the unsigned short that gives the number of
     *            values).
     * @param buf
     *            buffer to be used to call {@link #readUTF8 readUTF8},
     *            {@link #readClass(int,char[]) readClass} or {@link #readConst
     *            readConst}.
     * @param named
     *            if the annotation values are named or not.
     * @param av
     *            the visitor that must visit the values.
     * @return the end offset of the annotation values.
     */
    // row num 2006
    private int readAnnotationValues(int v, final char[] buf,
            final boolean named, final AnnotationVisitor av) {
        int i = readUnsignedShort(v);
        v += 2;
        if (named) {
            for (; i > 0; --i) {
                v = readAnnotationValue(v + 2, buf, readUTF8(v, buf), av);
            }
        } else {
            for (; i > 0; --i) {
                v = readAnnotationValue(v, buf, null, av);
            }
        }
        if (av != null) {
            av.visitEnd();
        }
        return v;
    }
......
    /**
     * Reads a value of an annotation and makes the given visitor visit it.
     * 
     * @param v
     *            the start offset in {@link #b b} of the value to be read
     *            (<i>not including the value name constant pool index</i>).
     * @param buf
     *            buffer to be used to call {@link #readUTF8 readUTF8},
     *            {@link #readClass(int,char[]) readClass} or {@link #readConst
     *            readConst}.
     * @param name
     *            the name of the value to be read.
     * @param av
     *            the visitor that must visit the value.
     * @return the end offset of the annotation value.
     */
    // row num 2041
    private int readAnnotationValue(int v, final char[] buf, final String name,
            final AnnotationVisitor av) {
        int i;
        if (av == null) {
            switch (b[v] & 0xFF) {
            case 'e': // enum_const_value
                return v + 5;
            case '@': // annotation_value
                return readAnnotationValues(v + 3, buf, true, null);
            case '[': // array_value
                return readAnnotationValues(v + 1, buf, false, null);
            default:
                return v + 3;
            }
        }
        switch (b[v++] & 0xFF) {
        case 'I': // pointer to CONSTANT_Integer
        case 'J': // pointer to CONSTANT_Long
        case 'F': // pointer to CONSTANT_Float
        case 'D': // pointer to CONSTANT_Double
            av.visit(name, readConst(readUnsignedShort(v), buf));
            v += 2;
            break;
        case 'B': // pointer to CONSTANT_Byte
            av.visit(name, (byte) readInt(items[readUnsignedShort(v)]));
            v += 2;
            break;
        case 'Z': // pointer to CONSTANT_Boolean
            av.visit(name,
                    readInt(items[readUnsignedShort(v)]) == 0 ? Boolean.FALSE
                            : Boolean.TRUE);
            v += 2;
            break;
        case 'S': // pointer to CONSTANT_Short
            av.visit(name, (short) readInt(items[readUnsignedShort(v)]));
            v += 2;
            break;
        case 'C': // pointer to CONSTANT_Char
            av.visit(name, (char) readInt(items[readUnsignedShort(v)]));
            v += 2;
            break;
        case 's': // pointer to CONSTANT_Utf8
            av.visit(name, readUTF8(v, buf));
            v += 2;
            break;
        case 'e': // enum_const_value
            av.visitEnum(name, readUTF8(v, buf), readUTF8(v + 2, buf));
            v += 4;
            break;
        case 'c': // class_info
            av.visit(name, Type.getType(readUTF8(v, buf)));
            v += 2;
            break;
        case '@': // annotation_value
            v = readAnnotationValues(v + 2, buf, true,
                    av.visitAnnotation(name, readUTF8(v, buf)));
            break;
        case '[': // array_value
            int size = readUnsignedShort(v);
            v += 2;
            if (size == 0) {
                return readAnnotationValues(v - 2, buf, false,
                        av.visitArray(name));
            }
            switch (this.b[v++] & 0xFF) {
            case 'B':
                byte[] bv = new byte[size];
                for (i = 0; i < size; i++) {
                    bv[i] = (byte) readInt(items[readUnsignedShort(v)]);
                    v += 3;
                }
                av.visit(name, bv);
                --v;
                break;
            case 'Z':
                boolean[] zv = new boolean[size];
                for (i = 0; i < size; i++) {
                    zv[i] = readInt(items[readUnsignedShort(v)]) != 0;
                    v += 3;
                }
                av.visit(name, zv);
                --v;
                break;
            case 'S':
                short[] sv = new short[size];
                for (i = 0; i < size; i++) {
                    sv[i] = (short) readInt(items[readUnsignedShort(v)]);
                    v += 3;
                }
                av.visit(name, sv);
                --v;
                break;
            case 'C':
                char[] cv = new char[size];
                for (i = 0; i < size; i++) {
                    cv[i] = (char) readInt(items[readUnsignedShort(v)]);
                    v += 3;
                }
                av.visit(name, cv);
                --v;
                break;
            case 'I':
                int[] iv = new int[size];
                for (i = 0; i < size; i++) {
                    iv[i] = readInt(items[readUnsignedShort(v)]);
                    v += 3;
                }
                av.visit(name, iv);
                --v;
                break;
            case 'J':
                long[] lv = new long[size];
                for (i = 0; i < size; i++) {
                    lv[i] = readLong(items[readUnsignedShort(v)]);
                    v += 3;
                }
                av.visit(name, lv);
                --v;
                break;
            case 'F':
                float[] fv = new float[size];
                for (i = 0; i < size; i++) {
                    fv[i] = Float
                            .intBitsToFloat(readInt(items[readUnsignedShort(v)]));
                    v += 3;
                }
                av.visit(name, fv);
                --v;
                break;
            case 'D':
                double[] dv = new double[size];
                for (i = 0; i < size; i++) {
                    dv[i] = Double
                            .longBitsToDouble(readLong(items[readUnsignedShort(v)]));
                    v += 3;
                }
                av.visit(name, dv);
                --v;
                break;
            default:
                v = readAnnotationValues(v - 3, buf, false, av.visitArray(name));
            }
        }
        return v;
    }

在2041行处的方法上，通过‘@’又重新进行迭代，即会将该类上本身的注解进行迭代查找，作为她的注解，进行注解的组合。