C++ variable declation syntax

问题:

I recently came across this construct: int(m); which seems to be equivalent to: int m;

Oddly, I have never seen this particular idiom before. Can someone point me to a reference where I can read the spec on this, or just explain directly? Does this also work in straight C?

Thanks, ConfusedDeveloper

回答1:

It is not an "idiom". It is just a redundant pair of parentheses. Grammatically, they can be there, but they serve no purpose.

Sometimes similar seemingly superfluous parentheses can be used to resolve ambiguity in C++ declarations, like

int a(int());

which declares a function can be turned into

int a((int()));

which is equivalent to

int a = int();

and defines a variable. But this is not exactly what you have in your case.

回答2:

It's also used to cast. Like,

double m= 10.0;
int n= int(m);

回答3:

In addition to other answers, sometimes declarator has to be parenthesized.
For example:

struct A {};

struct B { A a; };

namespace N {
  struct B { int a; };

  void f()
  {
    A (::B::*p) = &::B::a; // this () cannot be omitted
  }
}

If () is omitted in the above code, compiler recognizes a consecutive nested name specifier A::B instead of A and ::B, and will issue an error.
This parenthesis is necessary, but sometimes leads to a misleading situation.

struct A {
  int m;
  A() {}
  A( int ) {} // this isn't called
};

int i;

int main()
{
  A(i); // this is a declaration
  i.m = 1; // ok
}

In the above code, A(i) is a declaration(and also a definition in this case) of an object i, instead of a constructor call expression with an int argument i.

Hope this helps.