What are POD types in C++?

Question

I've come across this term POD-type a few times.
What does it mean?

Answer 1

POD stands for Plain Old Data - that is, a class (whether defined with the keyword struct or the keyword class) without constructors, destructors and virtual members functions. Wikipedia's article on POD goes into a bit more detail and defines it as:

A Plain Old Data Structure in C++ is an aggregate class that contains only PODS as members, has no user-defined destructor, no user-defined copy assignment operator, and no nonstatic members of pointer-to-member type.

Greater detail can be found in this answer for C++98/03. C++11 changed the rules surrounding POD, relaxing them greatly, thus necessitating a follow-up answer here.

Answer 2

Very informally:

A POD is a type (including classes) where the C++ compiler guarantees that there will be no "magic" going on in the structure: for example hidden pointers to vtables, offsets that get applied to the address when it is cast to other types (at least if the target's POD too), constructors, or destructors. Roughly speaking, a type is a POD when the only things in it are built-in types and combinations of them. The result is something that "acts like" a C type.

Less informally:

• int, char, wchar_t, bool, float, double are PODs, as are long/short and signed/unsigned versions of them.
• pointers (including pointer-to-function and pointer-to-member) are PODs,
• enums are PODs
• a const or volatile POD is a POD.
• a class, struct or union of PODs is a POD provided that all non-static data members are public, and it has no base class and no constructors, destructors, or virtual methods. Static members don't stop something being a POD under this rule. This rule has changed in C++11 and certain private members are allowed: Can a class with all private members be a POD class?
• Wikipedia is wrong to say that a POD cannot have members of type pointer-to-member. Or rather, it's correct for the C++98 wording, but TC1 made explicit that pointers-to-member are POD.

Formally (C++03 Standard):

3.9(10): "Arithmetic types (3.9.1), enumeration types, pointer types, and pointer to member types (3.9.2) and cv-qualified versions of these types (3.9.3) are collectively caller scalar types. Scalar types, POD-struct types, POD-union types (clause 9), arrays of such types and cv-qualified versions of these types (3.9.3) are collectively called POD types"

9(4): "A POD-struct is an aggregate class that has no non-static data members of type non-POD-struct, non-POD-union (or array of such types) or reference, and has no user-define copy operator and no user-defined destructor. Similarly a POD-union is an aggregate union that has no non-static data members of type non-POD-struct, non-POD-union (or array of such types) or reference, and has no user-define copy operator and no user-defined destructor.

8.5.1(1): "An aggregate is an array or class (clause 9) with no user-declared constructors (12.1), no private or protected non-static data members (clause 11), no base classes (clause 10) and no virtual functions (10.3)."

Answer 3

Plain Old Data

In short, it is all built-in data types (e.g. int, char, float, long, unsigned char, double, etc.) and all aggregation of POD data. Yes, it's a recursive definition. ;)

To be more clear, a POD is what we call "a struct": a unit or a group of units that just store data.

Answer 4

Why do we need to differentiate between POD's and non-POD's at all?

C++ started its life as an extension of C. While modern C++ is no longer a strict superset of C, people still expect a high level of compatibility between the two. The "C ABI" of a platform also frequently acts as a de-facto standard inter-language ABI for other languages on the platform.

Roughly speaking, a POD type is a type that is compatible with C and perhaps equally importantly is compatible with certain ABI optimisations.

To be compatible with C, we need to satisfy two constraints.

1. The layout must be the same as the corresponding C type.
2. The type must be passed to and returned from functions in the same way as the corresponding C type.

Certain C++ features are incompatible with this.

Virtual methods require the compiler to insert one or more pointers to virtual method tables, something that doesn't exist in C.

User-defined copy constructors, move constructors, copy assignments and destructors have implications for parameter passing and returning. Many C ABIs pass and return small parameters in registers, but the references passed to the user defined constructor/assigment/destructor can only work with memory locations.

So there is a need to define what types can be expected to be "C compatible" and what types cannot. C++03 was somewhat over-strict in this regard, any user-defined constructor would disable the built-in constructors and any attempt to add them back in would result in them being user-defined and hence the type being non-pod. C++11 opened things up quite a bit, by allowing the user to re-introduce the built-in constructors.

Answer 5

Examples of all non-POD cases with static_assert from C++11 to C++17 and POD effects

std::is_pod was added in C++11, so let's consider that standard onwards for now.

std::is_pod will be removed from C++20 as mentioned at https://stackoverflow.com/a/48435532/895245 , let's update this as support arrives for the replacements.

POD restrictions have become more and more relaxed as the standard evolved, I aim to cover all relaxations in the example through ifdefs.

libstdc++ has at tiny bit of testing at: https://github.com/gcc-mirror/gcc/blob/gcc-8_2_0-release/libstdc%2B%2B-v3/testsuite/20_util/is_pod/value.cc but it just too little. Maintainers: please merge this if you read this post. I'm lazy to check out all the C++ testsuite projects mentioned at: https://softwareengineering.stackexchange.com/questions/199708/is-there-a-compliance-test-for-c-compilers

#include <type_traits>
#include <array>
#include <vector>

int main() {
#if __cplusplus >= 201103L
    // # Not POD
    //
    // Non-POD examples. Let's just walk all non-recursive non-POD branches of cppreference.
    {
        // Non-trivial implies non-POD.
        // https://en.cppreference.com/w/cpp/named_req/TrivialType
        {
            // Has one or more default constructors, all of which are either
            // trivial or deleted, and at least one of which is not deleted.
            {
                // Not trivial because we removed the default constructor
                // by using our own custom non-default constructor.
                {
                    struct C {
                        C(int) {}
                    };
                    static_assert(std::is_trivially_copyable<C>(), "");
                    static_assert(!std::is_trivial<C>(), "");
                    static_assert(!std::is_pod<C>(), "");
                }

                // No, this is not a default trivial constructor either:
                // https://en.cppreference.com/w/cpp/language/default_constructor
                //
                // The constructor is not user-provided (i.e., is implicitly-defined or
                // defaulted on its first declaration)
                {
                    struct C {
                        C() {}
                    };
                    static_assert(std::is_trivially_copyable<C>(), "");
                    static_assert(!std::is_trivial<C>(), "");
                    static_assert(!std::is_pod<C>(), "");
                }
            }

            // Not trivial because not trivially copyable.
            {
                struct C {
                    C(C&) {}
                };
                static_assert(!std::is_trivially_copyable<C>(), "");
                static_assert(!std::is_trivial<C>(), "");
                static_assert(!std::is_pod<C>(), "");
            }
        }

        // Non-standard layout implies non-POD.
        // https://en.cppreference.com/w/cpp/named_req/StandardLayoutType
        {
            // Non static members with different access control.
            {
                // i is public and j is private.
                {
                    struct C {
                        public:
                            int i;
                        private:
                            int j;
                    };
                    static_assert(!std::is_standard_layout<C>(), "");
                    static_assert(!std::is_pod<C>(), "");
                }

                // These have the same access control.
                {
                    struct C {
                        private:
                            int i;
                            int j;
                    };
                    static_assert(std::is_standard_layout<C>(), "");
                    static_assert(std::is_pod<C>(), "");

                    struct D {
                        public:
                            int i;
                            int j;
                    };
                    static_assert(std::is_standard_layout<D>(), "");
                    static_assert(std::is_pod<D>(), "");
                }
            }

            // Virtual function.
            {
                struct C {
                    virtual void f() = 0;
                };
                static_assert(!std::is_standard_layout<C>(), "");
                static_assert(!std::is_pod<C>(), "");
            }

            // Non-static member that is reference.
            {
                struct C {
                    int &i;
                };
                static_assert(!std::is_standard_layout<C>(), "");
                static_assert(!std::is_pod<C>(), "");
            }

            // Neither:
            //
            // - has no base classes with non-static data members, or
            // - has no non-static data members in the most derived class
            //   and at most one base class with non-static data members
            {
                // Non POD because has two base classes with non-static data members.
                {
                    struct Base1 {
                        int i;
                    };
                    struct Base2 {
                        int j;
                    };
                    struct C : Base1, Base2 {};
                    static_assert(!std::is_standard_layout<C>(), "");
                    static_assert(!std::is_pod<C>(), "");
                }

                // POD: has just one base class with non-static member.
                {
                    struct Base1 {
                        int i;
                    };
                    struct C : Base1 {};
                    static_assert(std::is_standard_layout<C>(), "");
                    static_assert(std::is_pod<C>(), "");
                }

                // Just one base class with non-static member: Base1, Base2 has none.
                {
                    struct Base1 {
                        int i;
                    };
                    struct Base2 {};
                    struct C : Base1, Base2 {};
                    static_assert(std::is_standard_layout<C>(), "");
                    static_assert(std::is_pod<C>(), "");
                }
            }

            // Base classes of the same type as the first non-static data member.
            // TODO failing on GCC 8.1 -std=c++11, 14 and 17.
            {
                struct C {};
                struct D : C {
                    C c;
                };
                //static_assert(!std::is_standard_layout<C>(), "");
                //static_assert(!std::is_pod<C>(), "");
            };

            // C++14 standard layout new rules, yay!
            {
                // Has two (possibly indirect) base class subobjects of the same type.
                // Here C has two base classes which are indirectly "Base".
                //
                // TODO failing on GCC 8.1 -std=c++11, 14 and 17.
                // even though the example was copy pasted from cppreference.
                {
                    struct Q {};
                    struct S : Q { };
                    struct T : Q { };
                    struct U : S, T { };  // not a standard-layout class: two base class subobjects of type Q
                    //static_assert(!std::is_standard_layout<U>(), "");
                    //static_assert(!std::is_pod<U>(), "");
                }

                // Has all non-static data members and bit-fields declared in the same class
                // (either all in the derived or all in some base).
                {
                    struct Base { int i; };
                    struct Middle : Base {};
                    struct C : Middle { int j; };
                    static_assert(!std::is_standard_layout<C>(), "");
                    static_assert(!std::is_pod<C>(), "");
                }

                // None of the base class subobjects has the same type as
                // for non-union types, as the first non-static data member
                //
                // TODO: similar to the C++11 for which we could not make a proper example,
                // but with recursivity added.

                // TODO come up with an example that is POD in C++14 but not in C++11.
            }
        }
    }

    // # POD
    //
    // POD examples. Everything that does not fall neatly in the non-POD examples.
    {
        // Can't get more POD than this.
        {
            struct C {};
            static_assert(std::is_pod<C>(), "");
            static_assert(std::is_pod<int>(), "");
        }

        // Array of POD is POD.
        {
            struct C {};
            static_assert(std::is_pod<C>(), "");
            static_assert(std::is_pod<C[]>(), "");
        }

        // Private member: became POD in C++11
        // https://stackoverflow.com/questions/4762788/can-a-class-with-all-private-members-be-a-pod-class/4762944#4762944
        {
            struct C {
                private:
                    int i;
            };
#if __cplusplus >= 201103L
            static_assert(std::is_pod<C>(), "");
#else
            static_assert(!std::is_pod<C>(), "");
#endif
        }

        // Most standard library containers are not POD because they are not trivial,
        // which can be seen directly from their interface definition in the standard.
        // https://stackoverflow.com/questions/27165436/pod-implications-for-a-struct-which-holds-an-standard-library-container
        {
            static_assert(!std::is_pod<std::vector<int>>(), "");
            static_assert(!std::is_trivially_copyable<std::vector<int>>(), "");
            // Some might be though:
            // https://stackoverflow.com/questions/3674247/is-stdarrayt-s-guaranteed-to-be-pod-if-t-is-pod
            static_assert(std::is_pod<std::array<int, 1>>(), "");
        }
    }

    // # POD effects
    //
    // Now let's verify what effects does PODness have.
    //
    // Note that this is not easy to do automatically, since many of the
    // failures are undefined behaviour.
    //
    // A good initial list can be found at:
    // https://stackoverflow.com/questions/4178175/what-are-aggregates-and-pods-and-how-why-are-they-special/4178176#4178176
    {
        struct Pod {
            uint32_t i;
            uint64_t j;
        };
        static_assert(std::is_pod<Pod>(), "");

        struct NotPod {
            NotPod(uint32_t i, uint64_t j) : i(i), j(j) {}
            uint32_t i;
            uint64_t j;
        };
        static_assert(!std::is_pod<NotPod>(), "");

        // __attribute__((packed)) only works for POD, and is ignored for non-POD, and emits a warning
        // https://stackoverflow.com/questions/35152877/ignoring-packed-attribute-because-of-unpacked-non-pod-field/52986680#52986680
        {
            struct C {
                int i;
            };

            struct D : C {
                int j;
            };

            struct E {
                D d;
            } /*__attribute__((packed))*/;

            static_assert(std::is_pod<C>(), "");
            static_assert(!std::is_pod<D>(), "");
            static_assert(!std::is_pod<E>(), "");
        }
    }
#endif
}

GitHub upstream.

Tested with:

for std in 11 14 17; do echo $std; g++-8 -Wall -Werror -Wextra -pedantic -std=c++$std pod.cpp; done

on Ubuntu 18.04, GCC 8.2.0.

Answer 6

As I understand POD (PlainOldData) is just a raw data - it does not need:

• to be constructed,
• to be destroyed,
• to have custom operators.
• Must not have virtual functions,
• and must not override operators.

How to check if something is a POD? Well, there is a struct for that called std::is_pod:

namespace std {
// Could use is_standard_layout && is_trivial instead of the builtin.
template<typename _Tp>
  struct is_pod
  : public integral_constant<bool, __is_pod(_Tp)>
  { };
}

(From header type_traits)

Reference:

• http://en.cppreference.com/w/cpp/types/is_pod
• http://en.wikipedia.org/wiki/Plain_old_data_structure
• http://en.wikipedia.org/wiki/Plain_Old_C++_Object
• File type_traits

Answer 7

A POD (plain old data) object has one of these data types--a fundamental type, pointer, union, struct, array, or class--with no constructor. Conversely, a non-POD object is one for which a constructor exists. A POD object begins its lifetime when it obtains storage with the proper size for its type and its lifetime ends when the storage for the object is either reused or deallocated.

PlainOldData types also must not have any of:

• Virtual functions (either their own, or inherited)
• Virtual base classes (direct or indirect).

A looser definition of PlainOldData includes objects with constructors; but excludes those with virtual anything. The important issue with PlainOldData types is that they are non-polymorphic. Inheritance can be done with POD types, however it should only be done for ImplementationInheritance (code reuse) and not polymorphism/subtyping.

A common (though not strictly correct) definition is that a PlainOldData type is anything that doesn't have a VeeTable.

Answer 8

The concept of POD and the type trait std::is_pod will be deprecated in C++20. See this question for further information.

Answer 9

With C++, Plain Old Data doesn't just mean that things like int, char, etc are the only types used. Plain Old Data really means in practice that you can take a struct memcpy it from one location in memory to another and things will work exactly like you would expect (i.e. not blow up). This breaks if your class, or any class your class contains, has as a member that is a pointer or a reference or a class that has a virtual function. Essentially, if pointers have to be involved somewhere, its not Plain Old Data.