Reputation: 16193
Function argument names in declarations (that most likely reside in the header file) are seemingly completely ignored by the compiler. What are the reasons for allowing the following to compile using either declaration version 1 or 2?
implementation
void A::doStuff(int numElements, float* data)
{
//stuff
}
declaration - Version 1
class A
{
public:
void doStuff(int numElements, float* data);
}
declaration - Version 2
class A
{
public:
void doStuff(int, float*);
}
Upvotes: 6
Views: 859
Reputation: 24907
..because when the headers are included in other modules it only needs the types to generate the correct code. The names ae often useful and convenient, but nopt absolutely necessary.
Upvotes: 1
Reputation: 873
The compiler only needs to know what kind of arguments the method requires. It's unimportant for the compiler how you call them.
The compiler needs to know the argument types for several reasons:
However, I suggest to use the first header version. It helps other developers (and yourself) to use the functions and know what parameters have which meaning.
Upvotes: 11
Reputation: 258678
Parameter names aren't part of the function signature. Unless you use them, you don't need to have names even in the function implementation.
Upvotes: 6
Reputation: 272802
Because the names don't affect anything the compiler does outside the function.
Upvotes: 5
Reputation: 137442
The only reason I can think about that version 1 is better is readability. They are ignored as they don't matter for the compiler.
Upvotes: 1