序列化多态类型的常见混淆

Question

我看过许多涉及序列化派生 classes 的问题、教程和文档，但我未能就几个问题达成共识，包括（并在以下代码中说明）：

boost::serialization::base_object 对比 BOOST_SERIALIZATION_BASE_OBJECT_NVP
archive & mData; 对比 archive & BOOST_SERIALIZATION_NVP(mData);
BOOST_SERIALIZATION_ASSUME_ABSTRACT(AbstractPoint);
需要 serialize() 层次结构中的 class 不需要序列化任何内容。

代码：

#include <boost/archive/text_oarchive.hpp>
#include <boost/archive/text_iarchive.hpp>
#include <boost/serialization/shared_ptr.hpp>
#include <boost/serialization/base_object.hpp>

#include <fstream>

class AbstractPoint
{
public:
    virtual ~AbstractPoint(){}
    virtual void DoSomething() = 0;

    // Even though the class is abstract, we still need this
    template<class TArchive>
    void serialize(TArchive& archive, const unsigned int version)
    {
        // do nothing
    }
};

// This doesn't seem to do anything
//BOOST_SERIALIZATION_ASSUME_ABSTRACT(AbstractPoint);

class Point : public AbstractPoint
{
public:
    Point() = default;
    Point(const double data) : mData(data) {}

    void DoSomething(){}

    template<class TArchive>
    void serialize(TArchive& archive, const unsigned int version)
    {
        // These two seem equivalent. Without one of them, unregistered void cast
        archive & boost::serialization::base_object<AbstractPoint>(*this);
        //archive & BOOST_SERIALIZATION_BASE_OBJECT_NVP(AbstractPoint);

        // These two seem equivalent
        archive & mData;
        //archive & BOOST_SERIALIZATION_NVP(mData);
    }

    double mData;
};

int main()
{
    std::shared_ptr<AbstractPoint> point(new Point(7.4));

    std::ofstream outputStream("test.txt");
    boost::archive::text_oarchive outputArchive(outputStream);
    outputArchive.register_type<Point>();
    outputArchive << point;
    outputStream.close();

    std::shared_ptr<AbstractPoint> pointRead;
    std::ifstream inputStream("test.txt");
    boost::archive::text_iarchive inputArchive(inputStream);
    inputArchive.register_type<Point>();
    inputArchive >> pointRead;

    std::shared_ptr<Point> castedPoint = std::dynamic_pointer_cast<Point>(pointRead);
    std::cout << castedPoint->mData << std::endl;
    return 0;
}

另一个主要问题是在 "real" 环境中在哪里注册 classes（当有链接等时），但这似乎值得一个单独的问题。

最好在文档中有一个 "gold standard" 这类事情的例子，但至少在 Whosebug 上:)

Answer 1

boost::serialization::base_object 对比 BOOST_SERIALIZATION_BASE_OBJECT_NVP

只有具有元素命名的档案才需要 NVP 包装器，例如 XML。

除非您使用它，否则 base_object<> 更干净、更简单。

archive & mData; 对比 archive & BOOST_SERIALIZATION_NVP(mData);

同上

BOOST_SERIALIZATION_ASSUME_ABSTRACT(AbstractPoint);

我认为这只是一种优化 - 抑制每个存档类型的注册类型信息，因为你告诉框架它永远不会反序列化该类型的实例

需要 serialize() 层次结构中的 class 不需要序列化任何内容。

你不需要它，除非你需要那里关于多态基的类型信息。你什么时候需要那个？当你需要反序列化基类型的指针时。

因此，如果您

struct A{ virtual ~A(); };
struct B:A{};

struct C:B{};
struct D:B{};`

如果您（反）序列化 A*，您将需要序列化 A（而不是 B）。如果您（反）序列化 B*.

，则需要对 B 进行序列化

同样，如果您的类型不是多态的（虚拟的）或者您不这样使用它，则不需要任何基础序列化（例如，如果您（反）序列化 C 或 D直接）。

最后，如果你有 struct A{}; struct B:A{};，根本不需要告诉 Boost Serialization 关于基本类型，（你可以在 B 中进行序列化）。

根据您的示例进行更新：

case1.cpp 看起来不错

case2.cpp 当然需要调用基础序列化；不一定使用 base_object 因为你需要多态序列化：

template<class TArchive> void serialize(TArchive& archive, unsigned) {
    archive & boost::serialization::base_object<AbstractPoint>(*this)
            & mData;
    // OR:
    archive & static_cast<AbstractPoint&>(*this) 
            & mData;
    // OR even just:
    archive & mParentData 
            & mData;
}

case3.cpp：的确，它和案例1完全一样，但是有动态分配和对象跟踪

case4.cpp：和case1一模一样，但是有动态分配和对象跟踪； 注意!! 它需要显式序列化基础！

template<class TArchive> void serialize(TArchive& archive, unsigned) {
    archive & boost::serialization::base_object<AbstractPoint>(*this)
            & mData;
}

case5.cpp：是的，但更典型的是使用 boost/serialization/export.hpp

CLASS_EXPORT*

Bitrot insurance:

case1.cpp

case2.cpp

case3.cpp

case4.cpp

case5.cpp

Answer 2

按照@sehe 的建议，这里有一些使用示例：

序列化派生的 class 对象，不转发给父对象

#include <boost/archive/text_oarchive.hpp>
#include <boost/archive/text_iarchive.hpp>

#include <fstream>

class AbstractPoint
{
public:
    virtual ~AbstractPoint(){}
    virtual void DoSomething() = 0;
};

class Point : public AbstractPoint
{
public:
    Point() = default;
    Point(const double data) : mData(data) {}

    void DoSomething(){}

    template<class TArchive>
    void serialize(TArchive& archive, const unsigned int version)
    {
        archive & mData;
    }

    double mData;
};

int main()
{
    Point point(7.4);

    std::ofstream outputStream("test.txt");
    boost::archive::text_oarchive outputArchive(outputStream);
    outputArchive << point;
    outputStream.close();

    Point pointRead;
    std::ifstream inputStream("test.txt");
    boost::archive::text_iarchive inputArchive(inputStream);
    inputArchive >> pointRead;

    std::cout << pointRead.mData << std::endl;
    return 0;
}

序列化派生的 class 对象，包括（自动）转发给父对象：

#include <boost/archive/text_oarchive.hpp>
#include <boost/archive/text_iarchive.hpp>

#include <fstream>

class AbstractPoint
{
public:
    virtual ~AbstractPoint(){}
    virtual void DoSomething() = 0;

    double mParentData = 3.1;

    template<class TArchive>
    void serialize(TArchive& archive, const unsigned int version)
    {
        archive & mParentData;
    }
};

class Point : public AbstractPoint
{
public:
    Point() = default;
    Point(const double data) : mData(data) {}

    void DoSomething(){}

    template<class TArchive>
    void serialize(TArchive& archive, const unsigned int version)
    {
        // this is not required, the parent serialize() seems to be called automatically
        // archive & boost::serialization::base_object<AbstractPoint>(*this);

        archive & mData;
    }

    double mData;
};

int main()
{
    Point point(7.4);

    std::ofstream outputStream("test.txt");
    boost::archive::text_oarchive outputArchive(outputStream);
    outputArchive << point;
    outputStream.close();

    Point pointRead;
    std::ifstream inputStream("test.txt");
    boost::archive::text_iarchive inputArchive(inputStream);
    inputArchive >> pointRead;

    std::cout << pointRead.mParentData << std::endl;
    std::cout << pointRead.mData << std::endl;
    return 0;
}

序列化派生的class指针，不转发给父 （注意对象案例没有任何变化）

#include <boost/archive/text_oarchive.hpp>
#include <boost/archive/text_iarchive.hpp>
#include <boost/serialization/shared_ptr.hpp>

#include <fstream>

class AbstractPoint
{
public:
    virtual ~AbstractPoint(){}
    virtual void DoSomething() = 0;
};

class Point : public AbstractPoint
{
public:
    Point() = default;
    Point(const double data) : mData(data) {}

    void DoSomething(){}

    template<class TArchive>
    void serialize(TArchive& archive, const unsigned int version)
    {
        archive & mData;
    }

    double mData;
};

int main()
{
    std::shared_ptr<Point> point(new Point(7.4));

    std::ofstream outputStream("test.txt");
    boost::archive::text_oarchive outputArchive(outputStream);
    outputArchive << point;
    outputStream.close();

    std::shared_ptr<Point> pointRead;
    std::ifstream inputStream("test.txt");
    boost::archive::text_iarchive inputArchive(inputStream);
    inputArchive >> pointRead;

    std::cout << pointRead->mData << std::endl;
    return 0;
}

序列化派生的class指针，转发给父 （注意对象案例没有任何变化）

#include <boost/archive/text_oarchive.hpp>
#include <boost/archive/text_iarchive.hpp>
#include <boost/serialization/shared_ptr.hpp>

#include <fstream>

class AbstractPoint
{
public:
    virtual ~AbstractPoint(){}
    virtual void DoSomething() = 0;

    template<class TArchive>
    void serialize(TArchive& archive, const unsigned int version)
    {
        archive & mParentData;
    }

    double mParentData = 3.1;
};

class Point : public AbstractPoint
{
public:
    Point() = default;
    Point(const double data) : mData(data) {}

    void DoSomething(){}

    template<class TArchive>
    void serialize(TArchive& archive, const unsigned int version)
    {
        archive & mData;
    }

    double mData;
};

int main()
{
    std::shared_ptr<Point> point(new Point(7.4));

    std::ofstream outputStream("test.txt");
    boost::archive::text_oarchive outputArchive(outputStream);
    outputArchive << point;
    outputStream.close();

    std::shared_ptr<Point> pointRead;
    std::ifstream inputStream("test.txt");
    boost::archive::text_iarchive inputArchive(inputStream);
    inputArchive >> pointRead;

    std::cout << pointRead->mParentData << std::endl;
    std::cout << pointRead->mData << std::endl;
    return 0;
}

序列化基class指针 （我们现在必须在档案中注册派生的类型 class，以及使用 boost::serialization::base_object）

#include <boost/archive/text_oarchive.hpp>
#include <boost/archive/text_iarchive.hpp>
#include <boost/serialization/shared_ptr.hpp>
#include <boost/serialization/base_object.hpp>

#include <fstream>

class AbstractPoint
{
public:
    virtual ~AbstractPoint(){}
    virtual void DoSomething() = 0;

    // This is required if we want to serialize an AbstractPoint pointer
    template<class TArchive>
    void serialize(TArchive& archive, const unsigned int version)
    {
        // do nothing
    }
};

class Point : public AbstractPoint
{
public:
    Point() = default;
    Point(const double data) : mData(data) {}

    void DoSomething(){}

    template<class TArchive>
    void serialize(TArchive& archive, const unsigned int version)
    {
        // Without this, we get unregistered void cast
        archive & boost::serialization::base_object<AbstractPoint>(*this);

        archive & mData;
    }

    double mData;
};

int main()
{
    std::shared_ptr<AbstractPoint> point(new Point(7.4));

    std::ofstream outputStream("test.txt");
    boost::archive::text_oarchive outputArchive(outputStream);
    outputArchive.register_type<Point>();
    outputArchive << point;
    outputStream.close();

    std::shared_ptr<AbstractPoint> pointRead;
    std::ifstream inputStream("test.txt");
    boost::archive::text_iarchive inputArchive(inputStream);
    inputArchive.register_type<Point>();
    inputArchive >> pointRead;

    std::shared_ptr<Point> castedPoint = std::dynamic_pointer_cast<Point>(pointRead);
    std::cout << castedPoint->mData << std::endl;
    return 0;
}

序列化多态类型的常见混淆

Common confusions with serializing polymorphic types

c++

serialization

boost

boost-serialization

根据您的示例进行更新：