为什么 Gson.toJson 将通用字段序列化为空 JSON 对象？

我有一个包含 T 类型字段的泛型类，Gson 将此字段序列化为空对象。我在下面添加了代码来演示该问题。读回 JSON 似乎没问题（只要您提供正确的类型标记）。

import java.lang.reflect.Type;
import com.google.gson.Gson;
import com.google.gson.reflect.TypeToken;

public class GsonIssue {
    static class AbstractThing {
        private String fieldA = "valueA";

        public String getFieldA() {
            return fieldA;
        }

        public void setFieldA(String fieldA) {
            this.fieldA = fieldA;
        }

        @Override
        public String toString() {
            return "AbstractThing [fieldA=" + fieldA + "]";
        }
    }

    static class Thing extends AbstractThing {
        private String fieldB = "valueB";

        @Override
        public String toString() {
            return "Thing [fieldB=" + fieldB + ", fieldA=" + getFieldA() + "]";
        }
    }

    static class Wrapper<T extends AbstractThing> {
        private T abstractThing;
        private String standardField = "standard value";

        public Wrapper(T abstractThing) {
            this.abstractThing = abstractThing;
        }

        @Override
        public String toString() {
            return "Wrapper [abstractThing=" + abstractThing + ", standardField=" + standardField + "]";
        }
    }

    public static void main(String[] args) {
        Wrapper<Thing> wrapper = new Wrapper<Thing>(new Thing());

        Gson gson = new Gson();
        String json = gson.toJson(wrapper);
        System.out.println(json);
        // prints : {"abstractThing":{},"standardField":"standard value"}
        // so standardField is correctly serialized but not abstractThing.

        // but if we manually construct the expected json string, and parse it back, we get the expected object structure
        json = "{\"standardField\": \"some text\", " +
                "\"abstractThing\":{\"fieldB\" : \"arb value\", \"fieldA\" : \"another arb value\"}}";

        Type type = new TypeToken<Wrapper<Thing>>() {}.getType();
        Object fromJson = gson.fromJson(json, type);
        System.out.println(fromJson);
        // prints : Wrapper [abstractThing=Thing [fieldB=arb value, fieldA=another arb value], standardField=some text]
        // which is as expected
    }
}

来自他们的文档：

当您调用 toJson(obj) 时，Gson 会调用 obj.getClass() 来获取要序列化的字段的信息。同样，您通常可以在 fromJson(json, MyClass.class) 方法中传递 MyClass.class 对象。如果对象是非泛型类型，则此方法可以正常工作。但是，如果对象是泛型类型，则泛型类型信息会因为 Java 类型擦除而丢失

您可以通过为泛型类型指定正确的参数化类型来解决此问题。您可以使用 TypeToken 类来完成此操作。

他们给出了以下示例List<T>:

Type listType = new TypeToken<List<String>>() {}.getType();
gson.toJson(myStrings, listType);

所以对于你的代码，你需要......

Type myType = new TypeToken<Wrapper<Thing>>() {}.getType();
String json = gson.toJson(wrapper, myType);

https://sites.google.com/site/gson/gson-user-guide#TOC-Serializing-and-Deserializing-Gener https://sites.google.com/site/gson/gson-user-guide#TOC-Serializing-and-Deserializing-Gener