Gson int类型被转换成double问题解决(完美解决)

一、问题复现

1.1、BaseResponse类

class BaseResponse<T>(
    val code: Int = -1,
    val message: String? = null
) {
    var data: T? = null

    fun isSuccess(): Boolean {
        return code == 0
    }
}

1.2、不做任何处理，直接用Gson.fromJson解析

val json = "{\"code\":200,\"message\":\"成功\",\"data\":{\"ints\":200,\"doubles\":200.98,\"floats\":29.0986,\"longs\":29323627832875342,\"string\":\"字符串\"}}"
val baseResponse = Gson().fromJson(json, BaseResponse::class.java)
println(JsonUtils.toJson(baseResponse))

1.3、解析结果

I/System.out: {"code":200,"data":{"ints":200.0,"doubles":200.98,"floats":29.0986,"longs":2.9323627832875344E16,"string":"字符串"},"message":"成功"}

1.4、问题

Int类型的"ints"被转换成了double类型、Long类型的"longs"也被转换了

二、解决方案

2.1、旧的(网上千篇一律的复制方案,其实没有任何卵用)

网上千篇一律就是新建TypeAdapter接口的子类，手动转换。MyDataTypeAdapter 

package com.xxx.baseapp.net;

import com.google.gson.Gson;
import com.google.gson.GsonBuilder;
import com.google.gson.ToNumberPolicy;
import com.google.gson.TypeAdapter;
import com.google.gson.TypeAdapterFactory;
import com.google.gson.internal.LinkedTreeMap;
import com.google.gson.internal.bind.ObjectTypeAdapter;
import com.google.gson.reflect.TypeToken;
import com.google.gson.stream.JsonReader;
import com.google.gson.stream.JsonToken;
import com.google.gson.stream.JsonWriter;

import java.io.IOException;
import java.lang.reflect.Field;
import java.util.ArrayList;
import java.util.Collections;
import java.util.List;
import java.util.Map;


public final class MyDataTypeAdapter extends TypeAdapter<Object> {
    public static final TypeAdapterFactory FACTORY = new TypeAdapterFactory() {
        @SuppressWarnings("unchecked")
        @Override
        public <T> TypeAdapter<T> create(Gson gson, TypeToken<T> type) {
            if (type.getRawType() == Object.class) {
                return (TypeAdapter<T>) new MapTypeAdapter(gson);
            }
            return null;
        }
    };

    private final Gson gson;

    private MyDataTypeAdapter(Gson gson) {
        this.gson = gson;
    }

    @Override
    public Object read(JsonReader in) throws IOException {
        JsonToken token = in.peek();
        //判断字符串的实际类型
        switch (token) {
            case BEGIN_ARRAY:
                List<Object> list = new ArrayList<>();
                in.beginArray();
                while (in.hasNext()) {
                    list.add(read(in));
                }
                in.endArray();
                return list;

            case BEGIN_OBJECT:
                Map<String, Object> map = new LinkedTreeMap<>();
                in.beginObject();
                while (in.hasNext()) {
                    map.put(in.nextName(), read(in));
                }
                in.endObject();
                return map;
            case STRING:
                return in.nextString();
            case NUMBER:
                String s = in.nextString();
                if (s.contains(".")) {
                    return Double.valueOf(s);
                } else {
                    try {
                        return Integer.valueOf(s);
                    } catch (Exception e) {
                        return Long.valueOf(s);
                    }
                }
            case BOOLEAN:
                return in.nextBoolean();
            case NULL:
                in.nextNull();
                return null;
            default:
                throw new IllegalStateException();
        }
    }

    @Override
    public void write(JsonWriter out, Object value) throws IOException {
        if (value == null) {
            out.nullValue();
            return;
        }
        //noinspection unchecked
        TypeAdapter<Object> typeAdapter = (TypeAdapter<Object>) gson.getAdapter(value.getClass());
        if (typeAdapter instanceof ObjectTypeAdapter) {
            out.beginObject();
            out.endObject();
            return;
        }
        typeAdapter.write(out, value);
    }

    /**
     * 使用自定义工厂方法取代 Gson 实例中的工厂方法
     */
    public Gson getGson() {
        Gson gson = new GsonBuilder().create();
        try {
            Field factories = Gson.class.getDeclaredField("factories");
            factories.setAccessible(true);
            Object o = factories.get(gson);
            Class<?>[] declaredClasses = Collections.class.getDeclaredClasses();
            for (Class c : declaredClasses) {
                if ("java.util.Collections$UnmodifiableList".equals(c.getName())) {
                    Field listField = c.getDeclaredField("list");
                    listField.setAccessible(true);
                    List<TypeAdapterFactory> list = (List<TypeAdapterFactory>) listField.get(o);
                    int i = list.indexOf(ObjectTypeAdapter.getFactory(ToNumberPolicy.DOUBLE));
                    list.set(i, MyDataTypeAdapter.FACTORY);
                    break;
                }
            }
        } catch (Exception e) {
            e.printStackTrace();
        }
        return gson;
    }

} 

2.2、修改方案（看了源码就快要解决了）

2.2.1、分析

通过分析GsonBuilder的构造方法源码：

   /**
   * Creates a GsonBuilder instance that can be used to build Gson with various configuration
   * settings. GsonBuilder follows the builder pattern, and it is typically used by first
   * invoking various configuration methods to set desired options, and finally calling
   * {@link #create()}.
   */
  public GsonBuilder() {
  }

  /**
   * Constructs a GsonBuilder instance from a Gson instance. The newly constructed GsonBuilder
   * has the same configuration as the previously built Gson instance.
   *
   * @param gson the gson instance whose configuration should by applied to a new GsonBuilder.
   */
  GsonBuilder(Gson gson) {
    ......
    this.objectToNumberStrategy = gson.objectToNumberStrategy;
    this.numberToNumberStrategy = gson.numberToNumberStrategy;
    this.reflectionFilters.addAll(gson.reflectionFilters);
  }


  /**
   * Configures Gson to apply a specific number strategy during deserialization of {@link Object}.
   *
   * @param objectToNumberStrategy the actual object-to-number strategy
   * @return a reference to this {@code GsonBuilder} object to fulfill the "Builder" pattern
   * @see ToNumberPolicy#DOUBLE The default object-to-number strategy
   * @since 2.8.9
   */
  public GsonBuilder setObjectToNumberStrategy(ToNumberStrategy objectToNumberStrategy) {
    this.objectToNumberStrategy = Objects.requireNonNull(objectToNumberStrategy);
    return this;
  }

可以发现，在构造的时候默认了Object和Number类转换成数字类型的处理策略，可以看到，Object转换默认为ToNumberPolicy.DOUBLE的dobule的类型，因此不单单是int会被转换成double，Long类型等等数值类型都会默认被转换成dobule类型。

2.2.2、策略类

/**
 * An enumeration that defines two standard number reading strategies and a couple of
 * strategies to overcome some historical Gson limitations while deserializing numbers as
 * {@link Object} and {@link Number}.
 *
 * @see ToNumberStrategy
 * @since 2.8.9
 */
public enum ToNumberPolicy implements ToNumberStrategy {

  /**
   * Using this policy will ensure that numbers will be read as {@link Double} values.
   * This is the default strategy used during deserialization of numbers as {@link Object}.
   */
  DOUBLE {
    @Override public Double readNumber(JsonReader in) throws IOException {
      return in.nextDouble();
    }
  },

  /**
   * Using this policy will ensure that numbers will be read as a lazily parsed number backed
   * by a string. This is the default strategy used during deserialization of numbers as
   * {@link Number}.
   */
  LAZILY_PARSED_NUMBER {
    @Override public Number readNumber(JsonReader in) throws IOException {
      return new LazilyParsedNumber(in.nextString());
    }
  },

  /**
   * Using this policy will ensure that numbers will be read as {@link Long} or {@link Double}
   * values depending on how JSON numbers are represented: {@code Long} if the JSON number can
   * be parsed as a {@code Long} value, or otherwise {@code Double} if it can be parsed as a
   * {@code Double} value. If the parsed double-precision number results in a positive or negative
   * infinity ({@link Double#isInfinite()}) or a NaN ({@link Double#isNaN()}) value and the
   * {@code JsonReader} is not {@link JsonReader#isLenient() lenient}, a {@link MalformedJsonException}
   * is thrown.
   */
  LONG_OR_DOUBLE {
    @Override public Number readNumber(JsonReader in) throws IOException, JsonParseException {
      String value = in.nextString();
      try {
        return Long.parseLong(value);
      } catch (NumberFormatException longE) {
        try {
          Double d = Double.valueOf(value);
          if ((d.isInfinite() || d.isNaN()) && !in.isLenient()) {
            throw new MalformedJsonException("JSON forbids NaN and infinities: " + d + "; at path " + in.getPreviousPath());
          }
          return d;
        } catch (NumberFormatException doubleE) {
          throw new JsonParseException("Cannot parse " + value + "; at path " + in.getPreviousPath(), doubleE);
        }
      }
    }
  },

  /**
   * Using this policy will ensure that numbers will be read as numbers of arbitrary length
   * using {@link BigDecimal}.
   */
  BIG_DECIMAL {
    @Override public BigDecimal readNumber(JsonReader in) throws IOException {
      String value = in.nextString();
      try {
        return new BigDecimal(value);
      } catch (NumberFormatException e) {
        throw new JsonParseException("Cannot parse " + value + "; at path " + in.getPreviousPath(), e);
      }
    }
  }

}

通过上面的策略类源码，我们发现：LAZILY_PARSED_NUMBER 就是处理数值类型的枚举，查看LazilyParsedNumber源码，发现确实进行了细致的数值类型处理划分：

/**
 * This class holds a number value that is lazily converted to a specific number type
 *
 * @author Inderjeet Singh
 */
@SuppressWarnings("serial") // ignore warning about missing serialVersionUID
public final class LazilyParsedNumber extends Number {
  private final String value;

  /** @param value must not be null */
  public LazilyParsedNumber(String value) {
    this.value = value;
  }

  @Override
  public int intValue() {
    try {
      return Integer.parseInt(value);
    } catch (NumberFormatException e) {
      try {
        return (int) Long.parseLong(value);
      } catch (NumberFormatException nfe) {
        return new BigDecimal(value).intValue();
      }
    }
  }

  @Override
  public long longValue() {
    try {
      return Long.parseLong(value);
    } catch (NumberFormatException e) {
      return new BigDecimal(value).longValue();
    }
  }

  @Override
  public float floatValue() {
    return Float.parseFloat(value);
  }

  @Override
  public double doubleValue() {
    return Double.parseDouble(value);
  }

  @Override
  public String toString() {
    return value;
  }

  /**
   * If somebody is unlucky enough to have to serialize one of these, serialize
   * it as a BigDecimal so that they won't need Gson on the other side to
   * deserialize it.
   */
  private Object writeReplace() throws ObjectStreamException {
    return new BigDecimal(value);
  }

  private void readObject(ObjectInputStream in) throws IOException {
    // Don't permit directly deserializing this class; writeReplace() should have written a replacement
    throw new InvalidObjectException("Deserialization is unsupported");
  }

  @Override
  public int hashCode() {
    return value.hashCode();
  }

  @Override
  public boolean equals(Object obj) {
    if (this == obj) {
      return true;
    }
    if (obj instanceof LazilyParsedNumber) {
      LazilyParsedNumber other = (LazilyParsedNumber) obj;
      return value == other.value || value.equals(other.value);
    }
    return false;
  }
}

 2.2.2、最终解决方案

既然知道构造的时候设置了默认的 ToNumberPolicy.DOUBLE ，那修改起来就简单了，我们调用设置转换的策略方法即可：

fun getGson(): Gson {
    val gsonBuilder = GsonBuilder()
    gsonBuilder.setObjectToNumberStrategy(ToNumberPolicy.LAZILY_PARSED_NUMBER)
    return gsonBuilder.create()
}

调用修改策略方案后的getGson解析 

val json = "{\"code\":200,\"message\":\"成功\",\"data\":{\"ints\":200,\"doubles\":200.98,\"floats\":29.0986,\"longs\":29323627832875342,\"string\":\"字符串\"}}"
val baseResponse = getGson().fromJson(json, BaseResponse::class.java)
println(JsonUtils.toJson(baseResponse))

测试结果：

I/System.out: {"code":200,"data":{"ints":200,"doubles":200.98,"floats":29.0986,"longs":29323627832875342,"string":"字符串"},"message":"成功"}

至此完美解决！！！！

网上千篇一律的复制粘贴说这可以解决那个可以解决，结果然并卵。这个问题困扰了我2天，最后还是完美解决