I am not 100% positive about this, but this might be to prevent a syntax ambiguity. For example, consider the following class:
class BadTimes {
struct Overloaded;
int Overloaded; // Legal, but a very strange idea.
int confusing(Overloaded); // <-- This line
};
What does the indicated line mean? As written, this is a declaration of a member function named confusing
that accepts as a parameter an object of type Overloaded
(whose name isn't specified in the function declaration) and returns an int
. If C++11 were to allow initializers to use parentheses, this would be ambiguous, because it could also be a definition of a member of type int
named confusing
that is initialized to the value of the data member Overloaded
. (This is related to the current issue with the Most Vexing Parse.)
By requiring curly braces, this ambiguity is removed:
class BadTimes {
struct Overloaded;
int Overloaded; // Legal, but a very strange idea.
int confusing{Overloaded}; // <-- This line
};
Now, it's clear that confusing
is actually an int
initialized to the value of Overloaded
, because there's no way to read it as a function declaration.
Hope this helps!
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…