overloading operator errors

I am using GMP to process huge numbers. I need to overload +,-,/ and * operators to handle operations between double, mpz_t (gmp type) and char*. When I try to declare the operator, I have errors:

declararion: mpf_t& operator+(mpf_t& v1, double v2);

error:

functions.cpp:10: error: `__mpf_struct (& operator+(__mpf_struct (&)[1], double)
)[1]' must have an argument of class or enumerated type

Thanks any help!

That's a weird looking error statement. Did you copy the declaration and error statement from your compiler before posting to the board, or did you try to reproduce them from your personal memory when posting?

Is this:

mpf_t& operator+(mpf_t& v1, double v2);

declared as a member function within the mpf_t class or as a friend function within the class?

If it's a friend function then try renaming the parameters. If that doesn't work then post the full definition please.

You can only overload operator for which at least one of the arguments is a CLASS or reference to it. You're trying to overload a operator that receives a double and a pointer..

Code:

class A{};
A* operator+(const A* a,char){}
int main(){
	A a;
	&a+'h';
	return 0;
}

and the error

Code:

$ make
g++ -Wall -c ./main.cpp -o obj/./main.o
./main.cpp:24: error: `A* operator+(const A*, char)' must have an argument of cl
ass or enumerated type
./main.cpp: In function `int main()':
./main.cpp:27: warning: statement has no effect
./main.cpp:29:2: warning: no newline at end of file
make: *** [obj/./main.o] Error 1

It's cool when explanations are more straightforward than the question. Since the OP was passing a reference to a GMP, whatever that is, type, it would therefore seem that the compiler doesn't recognize the GMP type, mpz_t or mpf_t or whatever, as a valid type/class.

In case you didn't know.....GMP has a C++ class interface.

gg

Yes, it seems that the compiler does not recognize mpf_t as a struct, very strange indeed. Codeplug, I know that GMP has a C++ interface. The problem is that I was trying to use GMP in Windows. I have not found a compiled version of GMP that supports the C++ interface, and I was unable to compile full GMP under Cygwin. I'll move to Linux, the Windows was a restriction from the project.

Thanks for the replyies anyway. I am already compiling GMP under Linux to use its C++ interface.

Uhm... I forgot my real problem, sorry! The most important overload would be:

Code:

mpf_t operator+(char*, double);

What is impossible. The problem is that I have huge constants exported from a mathematical software into C code. Also, the code has operations with double, like pow. I can't represent the constants like any type, except char*.

I have made a parser that translate these huge constants into strings (not the std::string). The only solution I can think of, it is to overload the plus operator to std::string and double. Then I change my parser to produce the following output:

3987543756234

to

std::string("3987543756234 ")

Is this a good solution? Thanks any reply.