Question about One Definition Rule
Hello dear EE-ers,
There is a quote:
Please, provide example about this statement, especially regarding "in no even can there be more that one definition".
Thanks.
There is a quote:
In the entire program, an object or non-inline function cannot have more than one definition; if an object or function is used, it must have exactly one definition. You can declare an object or function that is never used, in which case you don't have to provide a definition. In no event can there be more than one definition.
Please, provide example about this statement, especially regarding "in no even can there be more that one definition".
Thanks.
Since 'definition' is used as a supplement to 'implementation', that statement speaks for itself. One thing is missing though, while multiple definitions (implementations) are an obvious problem, multiple declarations might hit as well.
Since you asked for examples:
Since you asked for examples:
//all fine, one declaration and one definition for each one
void print(const string& data);
void print(const string& data1, const string& data2);
void print(const string& data) {
cout << data << endl;
}
void print(const string& data1, const string& data2) {
cout << data << " " << data2 << endl;
}
// Houston, we have a problem
void print(const string& data);
void print(const string& data) {
cout << data << endl;
}
void print(const string& data) {
cerr << data << endl;
}
So, definition is more general term than implementation and in case of C programming language definition is reduced to implementation, correct?
@Akashay K
Thanks for ODR violation, could you please provide something that I can test to violate in gcc compiler?
Thanks for ODR violation, could you please provide something that I can test to violate in gcc compiler?
>>So, definition is more general term than implementation and in case of C programming language definition is reduced to
>>implementation, correct?
Indeed. I'd usually also go by the term 'implementation', but 'definition' here means the same and also covers instances and variables.
>>implementation, correct?
Indeed. I'd usually also go by the term 'implementation', but 'definition' here means the same and also covers instances and variables.
I have got it, if we say
Defining instances of class?
int i = 6; //integer i implements number 6
Defining instances of class?
//tu1.cpp
g++ tu1.cpp tu2.cpp
compiles like a charm.
struct X {
X(int);
X(int, int);
};
X::X(int = 0) { }
class D: public X {};
D d2;
int main(){}
struct X {
X(int);
X(int, int);
};
X::X(int = 0, int = 0) {}
class D: public X { };
g++ tu1.cpp tu2.cpp
compiles like a charm.
Yes, because exist independently in two different compilation units that don't cross-reference each other.
The problem about doing so is that it would defy the very purpose of your example. E.g.
// X.h
struct X {
X(int);
X(int, int);
// tu1.cpp
#include "X.h"
X::X(int = 0) { }
class D: public X {};
D d2;
#include "X.h"
X::X(int = 0) {int i = i;}
class D: public X { };
$ g++ tu1.cpp tu2.cpp
/tmp/cc9v8yr7.o:tu2.cpp:(.text+0x0): multiple definition of `X::X(int)'
/tmp/ccsPYguC.o:tu1.cpp:(.text+0x0): first defined here
/tmp/cc9v8yr7.o:tu2.cpp:(.text+0x0): multiple definition of `X::X(int)'
/tmp/ccsPYguC.o:tu1.cpp:(.text+0x0): first defined here
You have omitted some closed curly braces and semicolons :) But it's okay, I know you quite for long time :)
//tu1.cpp
//tu1.cpp
To keep the knowledge base clean (as sarabande suggested), I will accept my own answer also as a Solution :)
//tu1.cpp
#include "X.h"
X::X(int = 0) {};
class D: public X {};
D d2;
//tu1.cpp
#include "X.h"
X::X(int = 0) { int i = i;}
class D: public X {};
To keep the knowledge base clean (as sarabande suggested), I will accept my own answer also as a Solution :)
Premium Content
You need an Expert Office subscription to comment.Start Free Trial