gson-comments/gson/src/main/java/com/google/gson/internal/$Gson$Types.java

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

688 lines
24 KiB
Java
Raw Normal View History

/*
* Copyright (C) 2008 Google Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package com.google.gson.internal;
import static com.google.gson.internal.$Gson$Preconditions.checkArgument;
import static java.util.Objects.requireNonNull;
import java.io.Serializable;
import java.lang.reflect.Array;
import java.lang.reflect.GenericArrayType;
import java.lang.reflect.GenericDeclaration;
import java.lang.reflect.Modifier;
import java.lang.reflect.ParameterizedType;
import java.lang.reflect.Type;
import java.lang.reflect.TypeVariable;
import java.lang.reflect.WildcardType;
import java.util.Arrays;
import java.util.Collection;
import java.util.HashMap;
import java.util.Map;
import java.util.NoSuchElementException;
import java.util.Objects;
import java.util.Properties;
/**
* Static methods for working with types.
*
* @author Bob Lee
* @author Jesse Wilson
*/
public final class $Gson$Types {
static final Type[] EMPTY_TYPE_ARRAY = new Type[] {};
private $Gson$Types() {
throw new UnsupportedOperationException();
}
/**
* Returns a new parameterized type, applying {@code typeArguments} to {@code rawType} and
* enclosed by {@code ownerType}.
*
* @return a {@link java.io.Serializable serializable} parameterized type.
*/
public static ParameterizedType newParameterizedTypeWithOwner(
Type ownerType, Type rawType, Type... typeArguments) {
return new ParameterizedTypeImpl(ownerType, rawType, typeArguments);
}
/**
* Returns an array type whose elements are all instances of {@code componentType}.
*
* @return a {@link java.io.Serializable serializable} generic array type.
*/
public static GenericArrayType arrayOf(Type componentType) {
return new GenericArrayTypeImpl(componentType);
}
/**
* Returns a type that represents an unknown type that extends {@code bound}. For example, if
* {@code bound} is {@code CharSequence.class}, this returns {@code ? extends CharSequence}. If
* {@code bound} is {@code Object.class}, this returns {@code ?}, which is shorthand for {@code ?
* extends Object}.
*/
public static WildcardType subtypeOf(Type bound) {
Type[] upperBounds;
if (bound instanceof WildcardType) {
upperBounds = ((WildcardType) bound).getUpperBounds();
} else {
upperBounds = new Type[] {bound};
}
return new WildcardTypeImpl(upperBounds, EMPTY_TYPE_ARRAY);
}
/**
* Returns a type that represents an unknown supertype of {@code bound}. For example, if {@code
* bound} is {@code String.class}, this returns {@code ? super String}.
*/
public static WildcardType supertypeOf(Type bound) {
Type[] lowerBounds;
if (bound instanceof WildcardType) {
lowerBounds = ((WildcardType) bound).getLowerBounds();
} else {
lowerBounds = new Type[] {bound};
}
return new WildcardTypeImpl(new Type[] {Object.class}, lowerBounds);
}
/**
* Returns a type that is functionally equal but not necessarily equal according to {@link
* Object#equals(Object) Object.equals()}. The returned type is {@link java.io.Serializable}.
*/
public static Type canonicalize(Type type) {
if (type instanceof Class) {
Class<?> c = (Class<?>) type;
return c.isArray() ? new GenericArrayTypeImpl(canonicalize(c.getComponentType())) : c;
} else if (type instanceof ParameterizedType) {
ParameterizedType p = (ParameterizedType) type;
return new ParameterizedTypeImpl(
p.getOwnerType(), p.getRawType(), p.getActualTypeArguments());
} else if (type instanceof GenericArrayType) {
GenericArrayType g = (GenericArrayType) type;
return new GenericArrayTypeImpl(g.getGenericComponentType());
} else if (type instanceof WildcardType) {
WildcardType w = (WildcardType) type;
return new WildcardTypeImpl(w.getUpperBounds(), w.getLowerBounds());
} else {
// type is either serializable as-is or unsupported
return type;
}
}
public static Class<?> getRawType(Type type) {
if (type instanceof Class<?>) {
// type is a normal class.
return (Class<?>) type;
} else if (type instanceof ParameterizedType) {
ParameterizedType parameterizedType = (ParameterizedType) type;
// getRawType() returns Type instead of Class; that seems to be an API mistake,
// see https://bugs.openjdk.org/browse/JDK-8250659
Type rawType = parameterizedType.getRawType();
checkArgument(rawType instanceof Class);
return (Class<?>) rawType;
} else if (type instanceof GenericArrayType) {
Type componentType = ((GenericArrayType) type).getGenericComponentType();
return Array.newInstance(getRawType(componentType), 0).getClass();
} else if (type instanceof TypeVariable) {
// we could use the variable's bounds, but that won't work if there are multiple.
// having a raw type that's more general than necessary is okay
return Object.class;
} else if (type instanceof WildcardType) {
Improve `TypeToken` creation validation (#2072) * Add comments regarding multiple bounds of wildcard * Remove WildcardType check in getCollectionElementType The returned Type is never a wildcard due to the changes made to getSupertype by commit b1fb9ca9a1bea5440bc6a5b506ccf67236b08243. * Remove redundant getRawType call from MapTypeAdapterFactory getRawType(TypeToken.getType()) is the same as calling TypeToken.getRawType(). * Make TypeToken members private * Remove incorrect statement about TypeToken wildcards It is possible to use wildcards as part of the type argument, e.g.: `new TypeToken<List<? extends CharSequence>>() {}` * Only allow direct subclasses of TypeToken Previously subclasses of subclasses (...) of TypeToken were allowed which can behave incorrectly when retrieving the type argument, e.g.: class SubTypeToken<T> extends TypeToken<Integer> {} new SubTypeToken<String>() {}.getType() This returned `String` despite the class extending TypeToken<Integer>. * Throw exception when TypeToken captures type variable Due to type erasure the runtime type argument for a type variable is not available. Therefore there is no point in capturing a type variable and it might even give a false sense of type-safety. * Make $Gson$Types members private * Rename $Gson$Types.getGenericSupertype parameter Rename the method parameter to match the documentation of the method and to be similar to getSupertype(...). * Improve tests and handle raw TypeToken supertype better * Make some $Gson$Types members package-private again to prevent synthetic accessors * Remove TypeToken check for type variable As mentioned in review comments, there are cases during serialization where usage of the type variable is not so problematic (but still not ideal).
2022-04-19 17:20:58 +02:00
Type[] bounds = ((WildcardType) type).getUpperBounds();
// Currently the JLS only permits one bound for wildcards so using first bound is safe
assert bounds.length == 1;
return getRawType(bounds[0]);
} else {
String className = type == null ? "null" : type.getClass().getName();
throw new IllegalArgumentException(
"Expected a Class, ParameterizedType, or GenericArrayType, but <"
+ type
+ "> is of type "
+ className);
}
}
Improve `TypeToken` creation validation (#2072) * Add comments regarding multiple bounds of wildcard * Remove WildcardType check in getCollectionElementType The returned Type is never a wildcard due to the changes made to getSupertype by commit b1fb9ca9a1bea5440bc6a5b506ccf67236b08243. * Remove redundant getRawType call from MapTypeAdapterFactory getRawType(TypeToken.getType()) is the same as calling TypeToken.getRawType(). * Make TypeToken members private * Remove incorrect statement about TypeToken wildcards It is possible to use wildcards as part of the type argument, e.g.: `new TypeToken<List<? extends CharSequence>>() {}` * Only allow direct subclasses of TypeToken Previously subclasses of subclasses (...) of TypeToken were allowed which can behave incorrectly when retrieving the type argument, e.g.: class SubTypeToken<T> extends TypeToken<Integer> {} new SubTypeToken<String>() {}.getType() This returned `String` despite the class extending TypeToken<Integer>. * Throw exception when TypeToken captures type variable Due to type erasure the runtime type argument for a type variable is not available. Therefore there is no point in capturing a type variable and it might even give a false sense of type-safety. * Make $Gson$Types members private * Rename $Gson$Types.getGenericSupertype parameter Rename the method parameter to match the documentation of the method and to be similar to getSupertype(...). * Improve tests and handle raw TypeToken supertype better * Make some $Gson$Types members package-private again to prevent synthetic accessors * Remove TypeToken check for type variable As mentioned in review comments, there are cases during serialization where usage of the type variable is not so problematic (but still not ideal).
2022-04-19 17:20:58 +02:00
private static boolean equal(Object a, Object b) {
return Objects.equals(a, b);
}
/** Returns true if {@code a} and {@code b} are equal. */
public static boolean equals(Type a, Type b) {
if (a == b) {
// also handles (a == null && b == null)
return true;
} else if (a instanceof Class) {
// Class already specifies equals().
return a.equals(b);
} else if (a instanceof ParameterizedType) {
if (!(b instanceof ParameterizedType)) {
return false;
}
// TODO: save a .clone() call
ParameterizedType pa = (ParameterizedType) a;
ParameterizedType pb = (ParameterizedType) b;
return equal(pa.getOwnerType(), pb.getOwnerType())
&& pa.getRawType().equals(pb.getRawType())
&& Arrays.equals(pa.getActualTypeArguments(), pb.getActualTypeArguments());
} else if (a instanceof GenericArrayType) {
if (!(b instanceof GenericArrayType)) {
return false;
}
GenericArrayType ga = (GenericArrayType) a;
GenericArrayType gb = (GenericArrayType) b;
return equals(ga.getGenericComponentType(), gb.getGenericComponentType());
} else if (a instanceof WildcardType) {
if (!(b instanceof WildcardType)) {
return false;
}
WildcardType wa = (WildcardType) a;
WildcardType wb = (WildcardType) b;
return Arrays.equals(wa.getUpperBounds(), wb.getUpperBounds())
&& Arrays.equals(wa.getLowerBounds(), wb.getLowerBounds());
} else if (a instanceof TypeVariable) {
if (!(b instanceof TypeVariable)) {
return false;
}
2010-12-06 20:47:41 +01:00
TypeVariable<?> va = (TypeVariable<?>) a;
TypeVariable<?> vb = (TypeVariable<?>) b;
return Objects.equals(va.getGenericDeclaration(), vb.getGenericDeclaration())
&& va.getName().equals(vb.getName());
} else {
// This isn't a type we support. Could be a generic array type, wildcard type, etc.
return false;
}
}
public static String typeToString(Type type) {
2010-12-06 20:47:41 +01:00
return type instanceof Class ? ((Class<?>) type).getName() : type.toString();
}
/**
* Returns the generic supertype for {@code supertype}. For example, given a class {@code
* IntegerSet}, the result for when supertype is {@code Set.class} is {@code Set<Integer>} and the
* result when the supertype is {@code Collection.class} is {@code Collection<Integer>}.
*/
Improve `TypeToken` creation validation (#2072) * Add comments regarding multiple bounds of wildcard * Remove WildcardType check in getCollectionElementType The returned Type is never a wildcard due to the changes made to getSupertype by commit b1fb9ca9a1bea5440bc6a5b506ccf67236b08243. * Remove redundant getRawType call from MapTypeAdapterFactory getRawType(TypeToken.getType()) is the same as calling TypeToken.getRawType(). * Make TypeToken members private * Remove incorrect statement about TypeToken wildcards It is possible to use wildcards as part of the type argument, e.g.: `new TypeToken<List<? extends CharSequence>>() {}` * Only allow direct subclasses of TypeToken Previously subclasses of subclasses (...) of TypeToken were allowed which can behave incorrectly when retrieving the type argument, e.g.: class SubTypeToken<T> extends TypeToken<Integer> {} new SubTypeToken<String>() {}.getType() This returned `String` despite the class extending TypeToken<Integer>. * Throw exception when TypeToken captures type variable Due to type erasure the runtime type argument for a type variable is not available. Therefore there is no point in capturing a type variable and it might even give a false sense of type-safety. * Make $Gson$Types members private * Rename $Gson$Types.getGenericSupertype parameter Rename the method parameter to match the documentation of the method and to be similar to getSupertype(...). * Improve tests and handle raw TypeToken supertype better * Make some $Gson$Types members package-private again to prevent synthetic accessors * Remove TypeToken check for type variable As mentioned in review comments, there are cases during serialization where usage of the type variable is not so problematic (but still not ideal).
2022-04-19 17:20:58 +02:00
private static Type getGenericSupertype(Type context, Class<?> rawType, Class<?> supertype) {
if (supertype == rawType) {
return context;
}
// we skip searching through interfaces if unknown is an interface
Improve `TypeToken` creation validation (#2072) * Add comments regarding multiple bounds of wildcard * Remove WildcardType check in getCollectionElementType The returned Type is never a wildcard due to the changes made to getSupertype by commit b1fb9ca9a1bea5440bc6a5b506ccf67236b08243. * Remove redundant getRawType call from MapTypeAdapterFactory getRawType(TypeToken.getType()) is the same as calling TypeToken.getRawType(). * Make TypeToken members private * Remove incorrect statement about TypeToken wildcards It is possible to use wildcards as part of the type argument, e.g.: `new TypeToken<List<? extends CharSequence>>() {}` * Only allow direct subclasses of TypeToken Previously subclasses of subclasses (...) of TypeToken were allowed which can behave incorrectly when retrieving the type argument, e.g.: class SubTypeToken<T> extends TypeToken<Integer> {} new SubTypeToken<String>() {}.getType() This returned `String` despite the class extending TypeToken<Integer>. * Throw exception when TypeToken captures type variable Due to type erasure the runtime type argument for a type variable is not available. Therefore there is no point in capturing a type variable and it might even give a false sense of type-safety. * Make $Gson$Types members private * Rename $Gson$Types.getGenericSupertype parameter Rename the method parameter to match the documentation of the method and to be similar to getSupertype(...). * Improve tests and handle raw TypeToken supertype better * Make some $Gson$Types members package-private again to prevent synthetic accessors * Remove TypeToken check for type variable As mentioned in review comments, there are cases during serialization where usage of the type variable is not so problematic (but still not ideal).
2022-04-19 17:20:58 +02:00
if (supertype.isInterface()) {
2010-12-06 20:47:41 +01:00
Class<?>[] interfaces = rawType.getInterfaces();
for (int i = 0, length = interfaces.length; i < length; i++) {
Improve `TypeToken` creation validation (#2072) * Add comments regarding multiple bounds of wildcard * Remove WildcardType check in getCollectionElementType The returned Type is never a wildcard due to the changes made to getSupertype by commit b1fb9ca9a1bea5440bc6a5b506ccf67236b08243. * Remove redundant getRawType call from MapTypeAdapterFactory getRawType(TypeToken.getType()) is the same as calling TypeToken.getRawType(). * Make TypeToken members private * Remove incorrect statement about TypeToken wildcards It is possible to use wildcards as part of the type argument, e.g.: `new TypeToken<List<? extends CharSequence>>() {}` * Only allow direct subclasses of TypeToken Previously subclasses of subclasses (...) of TypeToken were allowed which can behave incorrectly when retrieving the type argument, e.g.: class SubTypeToken<T> extends TypeToken<Integer> {} new SubTypeToken<String>() {}.getType() This returned `String` despite the class extending TypeToken<Integer>. * Throw exception when TypeToken captures type variable Due to type erasure the runtime type argument for a type variable is not available. Therefore there is no point in capturing a type variable and it might even give a false sense of type-safety. * Make $Gson$Types members private * Rename $Gson$Types.getGenericSupertype parameter Rename the method parameter to match the documentation of the method and to be similar to getSupertype(...). * Improve tests and handle raw TypeToken supertype better * Make some $Gson$Types members package-private again to prevent synthetic accessors * Remove TypeToken check for type variable As mentioned in review comments, there are cases during serialization where usage of the type variable is not so problematic (but still not ideal).
2022-04-19 17:20:58 +02:00
if (interfaces[i] == supertype) {
return rawType.getGenericInterfaces()[i];
Improve `TypeToken` creation validation (#2072) * Add comments regarding multiple bounds of wildcard * Remove WildcardType check in getCollectionElementType The returned Type is never a wildcard due to the changes made to getSupertype by commit b1fb9ca9a1bea5440bc6a5b506ccf67236b08243. * Remove redundant getRawType call from MapTypeAdapterFactory getRawType(TypeToken.getType()) is the same as calling TypeToken.getRawType(). * Make TypeToken members private * Remove incorrect statement about TypeToken wildcards It is possible to use wildcards as part of the type argument, e.g.: `new TypeToken<List<? extends CharSequence>>() {}` * Only allow direct subclasses of TypeToken Previously subclasses of subclasses (...) of TypeToken were allowed which can behave incorrectly when retrieving the type argument, e.g.: class SubTypeToken<T> extends TypeToken<Integer> {} new SubTypeToken<String>() {}.getType() This returned `String` despite the class extending TypeToken<Integer>. * Throw exception when TypeToken captures type variable Due to type erasure the runtime type argument for a type variable is not available. Therefore there is no point in capturing a type variable and it might even give a false sense of type-safety. * Make $Gson$Types members private * Rename $Gson$Types.getGenericSupertype parameter Rename the method parameter to match the documentation of the method and to be similar to getSupertype(...). * Improve tests and handle raw TypeToken supertype better * Make some $Gson$Types members package-private again to prevent synthetic accessors * Remove TypeToken check for type variable As mentioned in review comments, there are cases during serialization where usage of the type variable is not so problematic (but still not ideal).
2022-04-19 17:20:58 +02:00
} else if (supertype.isAssignableFrom(interfaces[i])) {
return getGenericSupertype(rawType.getGenericInterfaces()[i], interfaces[i], supertype);
}
}
}
// check our supertypes
if (!rawType.isInterface()) {
while (rawType != Object.class) {
Class<?> rawSupertype = rawType.getSuperclass();
Improve `TypeToken` creation validation (#2072) * Add comments regarding multiple bounds of wildcard * Remove WildcardType check in getCollectionElementType The returned Type is never a wildcard due to the changes made to getSupertype by commit b1fb9ca9a1bea5440bc6a5b506ccf67236b08243. * Remove redundant getRawType call from MapTypeAdapterFactory getRawType(TypeToken.getType()) is the same as calling TypeToken.getRawType(). * Make TypeToken members private * Remove incorrect statement about TypeToken wildcards It is possible to use wildcards as part of the type argument, e.g.: `new TypeToken<List<? extends CharSequence>>() {}` * Only allow direct subclasses of TypeToken Previously subclasses of subclasses (...) of TypeToken were allowed which can behave incorrectly when retrieving the type argument, e.g.: class SubTypeToken<T> extends TypeToken<Integer> {} new SubTypeToken<String>() {}.getType() This returned `String` despite the class extending TypeToken<Integer>. * Throw exception when TypeToken captures type variable Due to type erasure the runtime type argument for a type variable is not available. Therefore there is no point in capturing a type variable and it might even give a false sense of type-safety. * Make $Gson$Types members private * Rename $Gson$Types.getGenericSupertype parameter Rename the method parameter to match the documentation of the method and to be similar to getSupertype(...). * Improve tests and handle raw TypeToken supertype better * Make some $Gson$Types members package-private again to prevent synthetic accessors * Remove TypeToken check for type variable As mentioned in review comments, there are cases during serialization where usage of the type variable is not so problematic (but still not ideal).
2022-04-19 17:20:58 +02:00
if (rawSupertype == supertype) {
return rawType.getGenericSuperclass();
Improve `TypeToken` creation validation (#2072) * Add comments regarding multiple bounds of wildcard * Remove WildcardType check in getCollectionElementType The returned Type is never a wildcard due to the changes made to getSupertype by commit b1fb9ca9a1bea5440bc6a5b506ccf67236b08243. * Remove redundant getRawType call from MapTypeAdapterFactory getRawType(TypeToken.getType()) is the same as calling TypeToken.getRawType(). * Make TypeToken members private * Remove incorrect statement about TypeToken wildcards It is possible to use wildcards as part of the type argument, e.g.: `new TypeToken<List<? extends CharSequence>>() {}` * Only allow direct subclasses of TypeToken Previously subclasses of subclasses (...) of TypeToken were allowed which can behave incorrectly when retrieving the type argument, e.g.: class SubTypeToken<T> extends TypeToken<Integer> {} new SubTypeToken<String>() {}.getType() This returned `String` despite the class extending TypeToken<Integer>. * Throw exception when TypeToken captures type variable Due to type erasure the runtime type argument for a type variable is not available. Therefore there is no point in capturing a type variable and it might even give a false sense of type-safety. * Make $Gson$Types members private * Rename $Gson$Types.getGenericSupertype parameter Rename the method parameter to match the documentation of the method and to be similar to getSupertype(...). * Improve tests and handle raw TypeToken supertype better * Make some $Gson$Types members package-private again to prevent synthetic accessors * Remove TypeToken check for type variable As mentioned in review comments, there are cases during serialization where usage of the type variable is not so problematic (but still not ideal).
2022-04-19 17:20:58 +02:00
} else if (supertype.isAssignableFrom(rawSupertype)) {
return getGenericSupertype(rawType.getGenericSuperclass(), rawSupertype, supertype);
}
rawType = rawSupertype;
}
}
// we can't resolve this further
Improve `TypeToken` creation validation (#2072) * Add comments regarding multiple bounds of wildcard * Remove WildcardType check in getCollectionElementType The returned Type is never a wildcard due to the changes made to getSupertype by commit b1fb9ca9a1bea5440bc6a5b506ccf67236b08243. * Remove redundant getRawType call from MapTypeAdapterFactory getRawType(TypeToken.getType()) is the same as calling TypeToken.getRawType(). * Make TypeToken members private * Remove incorrect statement about TypeToken wildcards It is possible to use wildcards as part of the type argument, e.g.: `new TypeToken<List<? extends CharSequence>>() {}` * Only allow direct subclasses of TypeToken Previously subclasses of subclasses (...) of TypeToken were allowed which can behave incorrectly when retrieving the type argument, e.g.: class SubTypeToken<T> extends TypeToken<Integer> {} new SubTypeToken<String>() {}.getType() This returned `String` despite the class extending TypeToken<Integer>. * Throw exception when TypeToken captures type variable Due to type erasure the runtime type argument for a type variable is not available. Therefore there is no point in capturing a type variable and it might even give a false sense of type-safety. * Make $Gson$Types members private * Rename $Gson$Types.getGenericSupertype parameter Rename the method parameter to match the documentation of the method and to be similar to getSupertype(...). * Improve tests and handle raw TypeToken supertype better * Make some $Gson$Types members package-private again to prevent synthetic accessors * Remove TypeToken check for type variable As mentioned in review comments, there are cases during serialization where usage of the type variable is not so problematic (but still not ideal).
2022-04-19 17:20:58 +02:00
return supertype;
}
/**
* Returns the generic form of {@code supertype}. For example, if this is {@code
* ArrayList<String>}, this returns {@code Iterable<String>} given the input {@code
* Iterable.class}.
*
* @param supertype a superclass of, or interface implemented by, this.
*/
Improve `TypeToken` creation validation (#2072) * Add comments regarding multiple bounds of wildcard * Remove WildcardType check in getCollectionElementType The returned Type is never a wildcard due to the changes made to getSupertype by commit b1fb9ca9a1bea5440bc6a5b506ccf67236b08243. * Remove redundant getRawType call from MapTypeAdapterFactory getRawType(TypeToken.getType()) is the same as calling TypeToken.getRawType(). * Make TypeToken members private * Remove incorrect statement about TypeToken wildcards It is possible to use wildcards as part of the type argument, e.g.: `new TypeToken<List<? extends CharSequence>>() {}` * Only allow direct subclasses of TypeToken Previously subclasses of subclasses (...) of TypeToken were allowed which can behave incorrectly when retrieving the type argument, e.g.: class SubTypeToken<T> extends TypeToken<Integer> {} new SubTypeToken<String>() {}.getType() This returned `String` despite the class extending TypeToken<Integer>. * Throw exception when TypeToken captures type variable Due to type erasure the runtime type argument for a type variable is not available. Therefore there is no point in capturing a type variable and it might even give a false sense of type-safety. * Make $Gson$Types members private * Rename $Gson$Types.getGenericSupertype parameter Rename the method parameter to match the documentation of the method and to be similar to getSupertype(...). * Improve tests and handle raw TypeToken supertype better * Make some $Gson$Types members package-private again to prevent synthetic accessors * Remove TypeToken check for type variable As mentioned in review comments, there are cases during serialization where usage of the type variable is not so problematic (but still not ideal).
2022-04-19 17:20:58 +02:00
private static Type getSupertype(Type context, Class<?> contextRawType, Class<?> supertype) {
if (context instanceof WildcardType) {
// Wildcards are useless for resolving supertypes. As the upper bound has the same raw type,
// use it instead
Improve `TypeToken` creation validation (#2072) * Add comments regarding multiple bounds of wildcard * Remove WildcardType check in getCollectionElementType The returned Type is never a wildcard due to the changes made to getSupertype by commit b1fb9ca9a1bea5440bc6a5b506ccf67236b08243. * Remove redundant getRawType call from MapTypeAdapterFactory getRawType(TypeToken.getType()) is the same as calling TypeToken.getRawType(). * Make TypeToken members private * Remove incorrect statement about TypeToken wildcards It is possible to use wildcards as part of the type argument, e.g.: `new TypeToken<List<? extends CharSequence>>() {}` * Only allow direct subclasses of TypeToken Previously subclasses of subclasses (...) of TypeToken were allowed which can behave incorrectly when retrieving the type argument, e.g.: class SubTypeToken<T> extends TypeToken<Integer> {} new SubTypeToken<String>() {}.getType() This returned `String` despite the class extending TypeToken<Integer>. * Throw exception when TypeToken captures type variable Due to type erasure the runtime type argument for a type variable is not available. Therefore there is no point in capturing a type variable and it might even give a false sense of type-safety. * Make $Gson$Types members private * Rename $Gson$Types.getGenericSupertype parameter Rename the method parameter to match the documentation of the method and to be similar to getSupertype(...). * Improve tests and handle raw TypeToken supertype better * Make some $Gson$Types members package-private again to prevent synthetic accessors * Remove TypeToken check for type variable As mentioned in review comments, there are cases during serialization where usage of the type variable is not so problematic (but still not ideal).
2022-04-19 17:20:58 +02:00
Type[] bounds = ((WildcardType) context).getUpperBounds();
// Currently the JLS only permits one bound for wildcards so using first bound is safe
assert bounds.length == 1;
context = bounds[0];
}
checkArgument(supertype.isAssignableFrom(contextRawType));
return resolve(
context,
contextRawType,
$Gson$Types.getGenericSupertype(context, contextRawType, supertype));
}
/**
* Returns the component type of this array type.
*
* @throws ClassCastException if this type is not an array.
*/
public static Type getArrayComponentType(Type array) {
return array instanceof GenericArrayType
? ((GenericArrayType) array).getGenericComponentType()
: ((Class<?>) array).getComponentType();
}
/**
* Returns the element type of this collection type.
*
* @throws IllegalArgumentException if this type is not a collection.
*/
public static Type getCollectionElementType(Type context, Class<?> contextRawType) {
Type collectionType = getSupertype(context, contextRawType, Collection.class);
if (collectionType instanceof ParameterizedType) {
return ((ParameterizedType) collectionType).getActualTypeArguments()[0];
}
return Object.class;
}
/**
* Returns a two element array containing this map's key and value types in positions 0 and 1
* respectively.
*/
public static Type[] getMapKeyAndValueTypes(Type context, Class<?> contextRawType) {
/*
* Work around a problem with the declaration of java.util.Properties. That
* class should extend Hashtable<String, String>, but it's declared to
* extend Hashtable<Object, Object>.
*/
if (context == Properties.class) {
return new Type[] {String.class, String.class}; // TODO: test subclasses of Properties!
}
Type mapType = getSupertype(context, contextRawType, Map.class);
// TODO: strip wildcards?
if (mapType instanceof ParameterizedType) {
ParameterizedType mapParameterizedType = (ParameterizedType) mapType;
return mapParameterizedType.getActualTypeArguments();
}
return new Type[] {Object.class, Object.class};
}
public static Type resolve(Type context, Class<?> contextRawType, Type toResolve) {
return resolve(context, contextRawType, toResolve, new HashMap<TypeVariable<?>, Type>());
}
private static Type resolve(
Type context,
Class<?> contextRawType,
Type toResolve,
Map<TypeVariable<?>, Type> visitedTypeVariables) {
// this implementation is made a little more complicated in an attempt to avoid object-creation
TypeVariable<?> resolving = null;
while (true) {
if (toResolve instanceof TypeVariable) {
2010-12-06 20:47:41 +01:00
TypeVariable<?> typeVariable = (TypeVariable<?>) toResolve;
Type previouslyResolved = visitedTypeVariables.get(typeVariable);
if (previouslyResolved != null) {
// cannot reduce due to infinite recursion
return (previouslyResolved == Void.TYPE) ? toResolve : previouslyResolved;
}
// Insert a placeholder to mark the fact that we are in the process of resolving this type
visitedTypeVariables.put(typeVariable, Void.TYPE);
if (resolving == null) {
resolving = typeVariable;
}
toResolve = resolveTypeVariable(context, contextRawType, typeVariable);
if (toResolve == typeVariable) {
break;
}
} else if (toResolve instanceof Class && ((Class<?>) toResolve).isArray()) {
Class<?> original = (Class<?>) toResolve;
Type componentType = original.getComponentType();
Type newComponentType =
resolve(context, contextRawType, componentType, visitedTypeVariables);
toResolve = equal(componentType, newComponentType) ? original : arrayOf(newComponentType);
break;
} else if (toResolve instanceof GenericArrayType) {
GenericArrayType original = (GenericArrayType) toResolve;
Type componentType = original.getGenericComponentType();
Type newComponentType =
resolve(context, contextRawType, componentType, visitedTypeVariables);
toResolve = equal(componentType, newComponentType) ? original : arrayOf(newComponentType);
break;
} else if (toResolve instanceof ParameterizedType) {
ParameterizedType original = (ParameterizedType) toResolve;
Type ownerType = original.getOwnerType();
Type newOwnerType = resolve(context, contextRawType, ownerType, visitedTypeVariables);
boolean changed = !equal(newOwnerType, ownerType);
Type[] args = original.getActualTypeArguments();
for (int t = 0, length = args.length; t < length; t++) {
Type resolvedTypeArgument =
resolve(context, contextRawType, args[t], visitedTypeVariables);
if (!equal(resolvedTypeArgument, args[t])) {
if (!changed) {
args = args.clone();
changed = true;
}
args[t] = resolvedTypeArgument;
}
}
toResolve =
changed
? newParameterizedTypeWithOwner(newOwnerType, original.getRawType(), args)
: original;
break;
} else if (toResolve instanceof WildcardType) {
WildcardType original = (WildcardType) toResolve;
Type[] originalLowerBound = original.getLowerBounds();
Type[] originalUpperBound = original.getUpperBounds();
if (originalLowerBound.length == 1) {
Type lowerBound =
resolve(context, contextRawType, originalLowerBound[0], visitedTypeVariables);
if (lowerBound != originalLowerBound[0]) {
toResolve = supertypeOf(lowerBound);
break;
}
} else if (originalUpperBound.length == 1) {
Type upperBound =
resolve(context, contextRawType, originalUpperBound[0], visitedTypeVariables);
if (upperBound != originalUpperBound[0]) {
toResolve = subtypeOf(upperBound);
break;
}
}
toResolve = original;
break;
} else {
break;
}
}
// ensure that any in-process resolution gets updated with the final result
if (resolving != null) {
visitedTypeVariables.put(resolving, toResolve);
}
return toResolve;
}
Improve `TypeToken` creation validation (#2072) * Add comments regarding multiple bounds of wildcard * Remove WildcardType check in getCollectionElementType The returned Type is never a wildcard due to the changes made to getSupertype by commit b1fb9ca9a1bea5440bc6a5b506ccf67236b08243. * Remove redundant getRawType call from MapTypeAdapterFactory getRawType(TypeToken.getType()) is the same as calling TypeToken.getRawType(). * Make TypeToken members private * Remove incorrect statement about TypeToken wildcards It is possible to use wildcards as part of the type argument, e.g.: `new TypeToken<List<? extends CharSequence>>() {}` * Only allow direct subclasses of TypeToken Previously subclasses of subclasses (...) of TypeToken were allowed which can behave incorrectly when retrieving the type argument, e.g.: class SubTypeToken<T> extends TypeToken<Integer> {} new SubTypeToken<String>() {}.getType() This returned `String` despite the class extending TypeToken<Integer>. * Throw exception when TypeToken captures type variable Due to type erasure the runtime type argument for a type variable is not available. Therefore there is no point in capturing a type variable and it might even give a false sense of type-safety. * Make $Gson$Types members private * Rename $Gson$Types.getGenericSupertype parameter Rename the method parameter to match the documentation of the method and to be similar to getSupertype(...). * Improve tests and handle raw TypeToken supertype better * Make some $Gson$Types members package-private again to prevent synthetic accessors * Remove TypeToken check for type variable As mentioned in review comments, there are cases during serialization where usage of the type variable is not so problematic (but still not ideal).
2022-04-19 17:20:58 +02:00
private static Type resolveTypeVariable(
Type context, Class<?> contextRawType, TypeVariable<?> unknown) {
Class<?> declaredByRaw = declaringClassOf(unknown);
// we can't reduce this further
if (declaredByRaw == null) {
return unknown;
}
Type declaredBy = getGenericSupertype(context, contextRawType, declaredByRaw);
if (declaredBy instanceof ParameterizedType) {
int index = indexOf(declaredByRaw.getTypeParameters(), unknown);
return ((ParameterizedType) declaredBy).getActualTypeArguments()[index];
}
return unknown;
}
private static int indexOf(Object[] array, Object toFind) {
for (int i = 0, length = array.length; i < length; i++) {
if (toFind.equals(array[i])) {
return i;
}
}
throw new NoSuchElementException();
}
/**
* Returns the declaring class of {@code typeVariable}, or {@code null} if it was not declared by
* a class.
*/
private static Class<?> declaringClassOf(TypeVariable<?> typeVariable) {
GenericDeclaration genericDeclaration = typeVariable.getGenericDeclaration();
return genericDeclaration instanceof Class ? (Class<?>) genericDeclaration : null;
}
static void checkNotPrimitive(Type type) {
2010-12-06 20:47:41 +01:00
checkArgument(!(type instanceof Class<?>) || !((Class<?>) type).isPrimitive());
}
/**
* Whether an {@linkplain ParameterizedType#getOwnerType() owner type} must be specified when
* constructing a {@link ParameterizedType} for {@code rawType}.
*
* <p>Note that this method might not require an owner type for all cases where Java reflection
* would create parameterized types with owner type.
*/
public static boolean requiresOwnerType(Type rawType) {
if (rawType instanceof Class<?>) {
Class<?> rawTypeAsClass = (Class<?>) rawType;
return !Modifier.isStatic(rawTypeAsClass.getModifiers())
&& rawTypeAsClass.getDeclaringClass() != null;
}
return false;
}
// Here and below we put @SuppressWarnings("serial") on fields of type `Type`. Recent Java
// compilers complain that the declared type is not Serializable. But in this context we go out of
// our way to ensure that the Type in question is either Class (which is serializable) or one of
// the nested Type implementations here (which are also serializable).
private static final class ParameterizedTypeImpl implements ParameterizedType, Serializable {
@SuppressWarnings("serial")
private final Type ownerType;
@SuppressWarnings("serial")
private final Type rawType;
@SuppressWarnings("serial")
private final Type[] typeArguments;
public ParameterizedTypeImpl(Type ownerType, Type rawType, Type... typeArguments) {
// TODO: Should this enforce that rawType is a Class? See JDK implementation of
// the ParameterizedType interface and https://bugs.openjdk.org/browse/JDK-8250659
requireNonNull(rawType);
if (ownerType == null && requiresOwnerType(rawType)) {
throw new IllegalArgumentException("Must specify owner type for " + rawType);
}
this.ownerType = ownerType == null ? null : canonicalize(ownerType);
this.rawType = canonicalize(rawType);
this.typeArguments = typeArguments.clone();
for (int t = 0, length = this.typeArguments.length; t < length; t++) {
requireNonNull(this.typeArguments[t]);
checkNotPrimitive(this.typeArguments[t]);
this.typeArguments[t] = canonicalize(this.typeArguments[t]);
}
}
@Override
public Type[] getActualTypeArguments() {
return typeArguments.clone();
}
@Override
public Type getRawType() {
return rawType;
}
@Override
public Type getOwnerType() {
return ownerType;
}
@Override
public boolean equals(Object other) {
return other instanceof ParameterizedType
&& $Gson$Types.equals(this, (ParameterizedType) other);
}
Improve `TypeToken` creation validation (#2072) * Add comments regarding multiple bounds of wildcard * Remove WildcardType check in getCollectionElementType The returned Type is never a wildcard due to the changes made to getSupertype by commit b1fb9ca9a1bea5440bc6a5b506ccf67236b08243. * Remove redundant getRawType call from MapTypeAdapterFactory getRawType(TypeToken.getType()) is the same as calling TypeToken.getRawType(). * Make TypeToken members private * Remove incorrect statement about TypeToken wildcards It is possible to use wildcards as part of the type argument, e.g.: `new TypeToken<List<? extends CharSequence>>() {}` * Only allow direct subclasses of TypeToken Previously subclasses of subclasses (...) of TypeToken were allowed which can behave incorrectly when retrieving the type argument, e.g.: class SubTypeToken<T> extends TypeToken<Integer> {} new SubTypeToken<String>() {}.getType() This returned `String` despite the class extending TypeToken<Integer>. * Throw exception when TypeToken captures type variable Due to type erasure the runtime type argument for a type variable is not available. Therefore there is no point in capturing a type variable and it might even give a false sense of type-safety. * Make $Gson$Types members private * Rename $Gson$Types.getGenericSupertype parameter Rename the method parameter to match the documentation of the method and to be similar to getSupertype(...). * Improve tests and handle raw TypeToken supertype better * Make some $Gson$Types members package-private again to prevent synthetic accessors * Remove TypeToken check for type variable As mentioned in review comments, there are cases during serialization where usage of the type variable is not so problematic (but still not ideal).
2022-04-19 17:20:58 +02:00
private static int hashCodeOrZero(Object o) {
return o != null ? o.hashCode() : 0;
}
@Override
public int hashCode() {
return Arrays.hashCode(typeArguments) ^ rawType.hashCode() ^ hashCodeOrZero(ownerType);
}
@Override
public String toString() {
int length = typeArguments.length;
if (length == 0) {
return typeToString(rawType);
}
StringBuilder stringBuilder = new StringBuilder(30 * (length + 1));
stringBuilder
.append(typeToString(rawType))
.append("<")
.append(typeToString(typeArguments[0]));
for (int i = 1; i < length; i++) {
stringBuilder.append(", ").append(typeToString(typeArguments[i]));
}
return stringBuilder.append(">").toString();
}
private static final long serialVersionUID = 0;
}
private static final class GenericArrayTypeImpl implements GenericArrayType, Serializable {
@SuppressWarnings("serial")
private final Type componentType;
public GenericArrayTypeImpl(Type componentType) {
requireNonNull(componentType);
this.componentType = canonicalize(componentType);
}
@Override
public Type getGenericComponentType() {
return componentType;
}
@Override
public boolean equals(Object o) {
return o instanceof GenericArrayType && $Gson$Types.equals(this, (GenericArrayType) o);
}
@Override
public int hashCode() {
return componentType.hashCode();
}
@Override
public String toString() {
return typeToString(componentType) + "[]";
}
private static final long serialVersionUID = 0;
}
/**
* The WildcardType interface supports multiple upper bounds and multiple lower bounds. We only
* support what the target Java version supports - at most one bound, see also
* https://bugs.openjdk.java.net/browse/JDK-8250660. If a lower bound is set, the upper bound must
* be Object.class.
*/
private static final class WildcardTypeImpl implements WildcardType, Serializable {
@SuppressWarnings("serial")
private final Type upperBound;
@SuppressWarnings("serial")
private final Type lowerBound;
public WildcardTypeImpl(Type[] upperBounds, Type[] lowerBounds) {
checkArgument(lowerBounds.length <= 1);
checkArgument(upperBounds.length == 1);
if (lowerBounds.length == 1) {
requireNonNull(lowerBounds[0]);
checkNotPrimitive(lowerBounds[0]);
checkArgument(upperBounds[0] == Object.class);
this.lowerBound = canonicalize(lowerBounds[0]);
this.upperBound = Object.class;
} else {
requireNonNull(upperBounds[0]);
checkNotPrimitive(upperBounds[0]);
this.lowerBound = null;
this.upperBound = canonicalize(upperBounds[0]);
}
}
@Override
public Type[] getUpperBounds() {
return new Type[] {upperBound};
}
@Override
public Type[] getLowerBounds() {
return lowerBound != null ? new Type[] {lowerBound} : EMPTY_TYPE_ARRAY;
}
@Override
public boolean equals(Object other) {
return other instanceof WildcardType && $Gson$Types.equals(this, (WildcardType) other);
}
@Override
public int hashCode() {
// this equals Arrays.hashCode(getLowerBounds()) ^ Arrays.hashCode(getUpperBounds());
return (lowerBound != null ? 31 + lowerBound.hashCode() : 1) ^ (31 + upperBound.hashCode());
}
@Override
public String toString() {
if (lowerBound != null) {
return "? super " + typeToString(lowerBound);
} else if (upperBound == Object.class) {
return "?";
} else {
return "? extends " + typeToString(upperBound);
}
}
private static final long serialVersionUID = 0;
}
}