I have the following code:
#include <iostream>
#include <vector>
namespace X {
std::ostream& operator<<(std::ostream& os,const std::vector<double>& v){
for (int i=0;i<v.size();i++){os << v[i] << " ";}
return os;
}
namespace Y {
struct A {std::vector<double> x;};
std::ostream& operator<<(std::ostream& os,const A& a){
os << a.x << std::endl;
return os;
}
}
}
using namespace X;
int main(int argc, char** argv) {
std::vector<double> v(10,0);
std::cout << v << std::endl;
Y::A a;
std::cout << a << std::endl;
return 0;
}
The first overload works, but the second one does not. For some reason it cannot find the first one. I get the error:
no match for 'operator<<' (operand types are 'std::ostream
{aka std::basic_ostream<char>}' and 'const std::vector<double>')
os << a.x << std::endl;
^
I do not understand why I get this error. For example something like this seems to be completely valid:
namespace A {
void foo(){}
namespace B {
void bar(){foo();}
}
}
However, the only way to fix the above problem was to put the second overload also in X. Why is it not possible to have it in the same namespace as the struct (ie. X::Y)?
PS: I was reading on ADL and I found some related questions (e.g. this and this, but what I understood from reading this, the above should work.
These operators cannot be overloaded because if we overload them it will make serious programming issues. For an example the sizeof operator returns the size of the object or datatype as an operand. This is evaluated by the compiler.
You can redefine or overload the function of most built-in operators in C++. These operators can be overloaded globally or on a class-by-class basis. Overloaded operators are implemented as functions and can be member functions or global functions.
Rules for Operator Overloading:You can only overload existing operators. You can't overload new operators. Some operators cannot be overloaded using a friend function. However, such operators can be overloaded using member function.
Input/Output Operators Overloading in C++ C++ is able to input and output the built-in data types using the stream extraction operator >> and the stream insertion operator <<. The stream insertion and stream extraction operators also can be overloaded to perform input and output for user-defined types like an object.
In Argument Depended Lookup (or Koenig Lookup), compiler adds to the scope of visibility all symbols declared in parent scopes of each parameter.
Even if Y
is "child namespace" of X
, they are not related in terms of ADL
. First of your parameters is type defined in std::
namespace, while second is local symbol (defined in the same namespace as the function itself).
Note, that because of reasons mentioned above, you will most likely get another error in this line:
std::cout << v << std::endl;
when compiler will not be able to find operator<<
overloaded for std::vector<double>
(because it lies inside namespace X
).
To solve this, you can use:
using X::operator<<
inside namespace Y
or move that overload.
If you are wondering, why foobar
example works: that's because ADL
(Argument Dependent Lookup) is about scope of parameters of functions, not functions themselves. In foobar
code, ADL
isn't applied.
As per other answers, I eventually deduced that ADL of operator<< was being impeded by the fact that it was taking place inside another operator<<.
Today's lesson: always write an operator<< overload in terms of a writer method :)
Here's the fix:
#include <iostream>
#include <vector>
namespace X
{
std::ostream& operator<<(std::ostream& os,const std::vector<double>& v){
for (int i=0;i<v.size();i++){os << v[i] << " ";}
return os;
}
namespace Y
{
struct A
{
std::vector<double> x;
void write(std::ostream&os) const {
os << x << std::endl;
}
};
std::ostream& operator<<(std::ostream& os,const A& a)
{
a.write(os);
return os;
}
}
}
using namespace X;
int main(int argc, char** argv)
{
std::vector<double> v(10,0);
std::cout << v << std::endl;
X::Y::A a;
std::cout << a << std::endl;
return 0;
}
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With