C++ custom data types

SolarNexus

I've been messin' with c++ for far too long now without actually doing anything, so I've decided to attempt something I've been considering for ages - a C++ implementation of a bittorrent client (big task, I know..).

So heres the problem - I want to create a custom data type, bint (bencoded int), its basically an integer converted to a string with "i" prefixed and "e" postfixed" (EG: int 5 = 5 whereas bint 5 = "i5e"). Thus far the only way I can think of is using a class - I tried using a struct but it just wont work the way I tried and I'm not entirely happy with using a class.

kdouglas

afaik your only option is a struct or class

happydude13

I'm not sure that what you have described makes sense...

Do you want a datatype that thinks it's an int ?
and a string ?

so you can use it as follows

bint t1 = 5;
bint t2 = "i5e";

and t1 == t2 ?

eof

SolarNexus

happydude13 wrote:

I'm not sure that what you have described makes sense...

Do you want a datatype that thinks it's an int ?
and a string ?

so you can use it as follows

bint t1 = 5;
bint t2 = "i5e";

and t1 == t2 ?

eof

sort of, but not quite.

I want a datatype which I can say something like mybint += 5 and still be valid, whilst also being able to get its bencoded string; think of it as a special string type that accepts arthimetic operators. I guess I'm just asking too much to get this simply, a class might be the best option.


an example:

bint b1 = 5;

while( true )
{
b1++;
cout << "b1: " << b1; // outputs a string similar to "b1: i123e"
}

satchmo

You'll want to use a class with operator overloading, that'll do it. If you create a constructor that takes a single int as a parameter (and isn't an explicit constructor), you should be able to do bint b1=5; and the compiler will fill in the rest. You'll also need to implement all the other operators... +, -, ++, --, etc.

And for the stream output you can also overload the global stream operator like this:

std::ostream& operator<<(std::ostream &str, const bint &b){
    str<<"i"<<b.intValue()<<"e";
    return str;
}

presuming your bint class has an int intValue(void) function.

SolarNexus

class bint
{
private:
	// data members
	int val;
	string benc;

	// computers
	void compute()
	{
		// numbers with leading 0's are invalid , fix later //
		//////////////////////////////////////////////////////

		// data members
		char buff[17];
		string pre = "i" , post = "e";

		// error check
		if( val == -0 )
			val = 0;

		// doit
		_itoa_s( val , buff , 10 );

		// return
		benc = ( pre + buff + post );
	}

public:

	// func
	bint(int);

	// mutators
	void setVal( int value )
	{
		val = value;
		compute();
	}
	
	// aggregators
	int getVal()
	{
		return val;
	}
	string getBencode()
	{
		return benc;
	}


};
bint::bint( int value )
{
	if( value == NULL )
	{
		val = 0;
		benc = "i0e";
	}
	else
	{
		val = value;
		compute();
	}
}

^^ thats what I finished off just after my previous post.

Tobias Greeshman

if(val == -0)
   val = 0 ;
 
// should this not be
if ( val < 0 )
   val = 0 ;

How can you get leading zero's?? val is an int not a string, leading zero's can't happen.

You just need some operator overloaded functions for operations like ==, !=, ++, --, +, -, =

SolarNexus

silas wrote:

if(val == -0)
   val = 0 ;
 
// should this not be
if ( val < 0 )
   val = 0 ;

How can you get leading zero's?? val is an int not a string, leading zero's can't happen.

You just need some operator overloaded functions for operations like ==, !=, ++, --, +, -, =

the reason for if( val == -0 ) is that -0 is not permitted in bencoding (http://www.bittorrent.com/protocol.html), but -1234 etc. is.

you can get leading 0's in an integer: 00234 is equal to 234