Help

[ladan.n]

Hi
i have a question:
how can we wirte a program that can print the 2 first bytes of a double ??

[Brad0407]

Can you clearify that, maybe give an example. And then try it on your own so we have some code work with. No one wants to write the program for you.

[ladan.n]

I don't want anyone to write a program for me !
i just needed help thats all !

[telepatico]

What about Mod division or something similar?

[Dino]

You can cast the field as something else to print the first 2 bytes, either with a (cast) operator or perhaps code up a union to redefine it.

Todd

[MacGyver]

Depends what you mean by the first two bytes. Doubles are generally not guarenteed to be kept in the same binary format that an integer is. You also run into issues with endianness.

If you don't care about any of that and just want to print it for the heck of it, then Todd's suggestion of a union is probably a good idea. The same can manually be done without a union, but a union takes care of some pointer and casting details.

[Elysia]

The easiest idea I find (if you don't need to worry about endianess) is to cast it to char*.

[MacGyver]

Here's something with the union idea:

#include <stdio.h>
#include <string.h>

typedef union
{
	char	asChar;
	short	asShort;
	int	asInt;
	long	asLong;
	double	asDouble;
}doublewrapper;

void dumpWrapper(const doublewrapper * const dw);

int main(void)
{
	doublewrapper dw;
	memset(&dw, 0, sizeof(doublewrapper));

	printf("Enter a double: ");
	fflush(stdout);

	scanf("%lf", &dw.asDouble);

	dumpWrapper(&dw);

	return 0;
}

void dumpWrapper(const doublewrapper * const dw)
{
	printf("Char\t\tShort\t\tInt\t\tLong\t\tDouble\n\n");
	printf("%08d\t%08d\t%08d\t%08ld\t%f\n", dw->asChar, dw->asShort, dw->asInt, dw->asLong, dw->asDouble);
	printf("%08X\t%08X\t%08X\t%08lX\t%f\n", dw->asChar, dw->asShort, dw->asInt, dw->asLong, dw->asDouble);
}

Disclaimer: This is very messy, non-portable (in the sense that you may receive different numerical results on different systems), and possibly has a few errors in it (since I rushed it).

[Elysia]

int main()
{
	int i = 0;
	double myd = 1234.0;
	char* p = (char*)&myd;
	for (; i < sizeof(myd); i++, p++)
	{
		printf("%i", (int)*p);
	}
	return 0;
}

Should do the same thing, I believe. Not tested, of course.

[dwks]

It might be a good idea to use unsigned chars, because ordinary chars can be signed, and the values dealt with might easily be negative when represented as signed chars. Unless you don't mind negative byte values, in which case you might want to use signed chars specifically, because ordinary chars can be signed or unsigned.

[ssharish2005]

bit masking could do as well i suppose

ssharish

[ladan.n]

thanks Guys!

[matsp]

And of course without knowing at least the byteorder of the architecture, you have no idea if you are displaying the least significant digits or the exponent and sign bytes of the double.

The whole excercise is pretty meaningless unless it's for the purpose of understanding unions/pointer casts.

--
Mats