Default equality operator for an empty union

Question

C++ allows us to make union class type without any data members. (cppreference even contain an example with it.)

Can one request the compiler to provide default implementation of equality comparison operator for such types? (It can be necessary for some generic code requiring comparison of objects.)

This program:

union U {
    constexpr bool operator ==(const U&) const = default;
};

// error in MSVC
bool a = ( U{} == U{} );

// error in Clang
constexpr bool b = ( U{} == U{} );

makes the compilers diverge:

Visual Studio's compiler implicitly deletes operator ==:

error C2280: 'bool U::operator ==(const U &) const': attempting to reference a deleted function
note: see declaration of 'U::operator =='
note: 'bool U::operator ==(const U &) const': function was implicitly deleted because 'U' is a union-like class
error C2088: built-in operator '==' cannot be applied to an operand of type 'U'

Clang defines the operator, but it does not permit using it in constant expressions:

note: pointer to temporary is not a constant expression
note: temporary created here
    9 | constexpr bool b = ( U{} == U{} );

GCC and EDG do not see any error and accept the program. Online demo: https://gcc.godbolt.org/z/WdqGYG4r1

Which implementation is correct here?

Why would you use this though? How is an empty union helpful? — Daniel
– Daniel, Commented Oct 18 at 19:16
@Daniel, empty union is very similar to empty struct, but it has some interesting properties like it cannot be a base class, which is something one wants to prevent sometimes. — Fedor
– Fedor, Commented Oct 18 at 20:06

Remy Lebeau · Accepted Answer · 2025-10-17 19:43:43Z

24

The wording from [class.compare.default] reads:

A defaulted <=> or == operator function for class C is defined as deleted if any non-static data member of C is of reference type or C has variant members ([class.union.anon]).

An empty union doesn't actually have variant members, so this restriction technically doesn't apply. So, I see no reason per the current wording why it shouldn't be valid — and once it's valid, it should certainly be usable as a constant expression. The relevant equality wording is defined in terms of subobjects, which there aren't any, so it should just be trivially true.

This is separate from the question of whether it should be valid. Probably not?

edited Oct 17 at 19:43

Remy Lebeau

609k36 gold badges516 silver badges875 bronze badges

answered Oct 17 at 19:37

Barry

312k32 gold badges733 silver badges1.1k bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Default equality operator for an empty union

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related