I'm not fully comprehending variants' usage, maybe someone can clear up what I'm doing wrong (possibly my approach).
Let the variant object be _Types of two classes, both inheriting the same class.
class base
{
public:
int foo;
};
class a: public base
{
int bar;
};
class b: public base
{
float bar;
};
byte variant_id = 0; // 0 for class A, 1 for class B
std::variant< base, a, b > variant;
Here's how I use the variant:
void func( )
{
auto& _variant = std::get< base >( variant ); // ideally would be = variant_id ? std::get< b >( variant ) : std::get< a >( variant )
_variant.foo = 20;
if ( variant_id == 1 )
{
auto& variant_ = std::get< b >( variant );
variant_.bar = 20.f;
}
else
{
auto& variant_ = std::get< a >( variant );
variant_.bar = 20;
}
Maybe a union is more effective?
union
{
a _a;
b _b;
} variant;
byte variant_id = 0;
Try not to query the variant for what type it holds. If you do, your code is essentially equivalent to a bunch of dynamic casts in an if chain, a code smell.
Instead, let the variant dispatch for you. And if you want to access the common base
of a
and b
, you do not need a base
member in that variant.
Use a visitor
std::variant< a, b > var;
void func( )
{
std::visit([](auto&& v) {
v.foo = 20; // Both `a` and `b` have a foo, this is well formed.
v.bar = 20; // Both have a `bar` that can be assigned a 20. This is well formed too
}, var);
}