Thread: Adding a Large number digit by digit

Originally Posted by Salem

Or just invert the test for even more simplicity, and perhaps a bit more obvious as to what it's about

Code:

if ( (number1[y] != 0) && (number2[y] != 0) ) {
    sum += carry;
    sum += number1[y] - '0' + number2[y] - '0';
    carry = sum / 10;
    result[y] = sum % 10 + '0';
    sum = 0;
}

True, I just prefer code where error/rare conditions are handled inside if statements. To me, that makes it clearer that the if statement handles abnormal conditions, and the rest of the code is free to execute and not dependent on an error *not* being true. A style difference, I guess

Originally Posted by matsp

In some cases, using the right way around of *p++ or *--p would perhaps cause better code on for example PDP-11, VAX or 68K, as those processors have "mov rA,(rB)+" and "mov rA,-(rB)", where the register rB is incremented/decremened automatically "at the same time as the access to memory" - but that would be specific to certain architectures, and definitely doesn't apply to the above case you described.

I stand corrected. Apologies.

QuantumPete

Why not create a new string of the correct size, fill it with however many zeros you need for padding, and copy the old string after that?

Or you might be able to do something like this, reusing the same string:

Code:

/* where:
    padding = number of zeros that need to be added to the string
    digits = number of digits currently in the string
    buffer = the number that needs to be padded --
        must have enough space for digits+padding digits plus a NULL
*/
memmove(buffer + padding, buffer, digits);
memset(buffer, '0', padding);

Code:

#include "stdafx.h"

That's non-standard, and you don't use anything from it. Your program should be able to compile without you including it. Try it.

Code:

unsigned char number1 [256] = {'0'};
unsigned char number2 [256] = {'0'};
unsigned char result [256] = {0};
unsigned char result1 [256] = {0};

When a compound initializer is missing some values, the remaining elements are filled with 0. This is why

Code:

int array[size] = {0};

fills every element in array with zeros. However, this code

Code:

char string[size] = {'0'};

will put '0' in the first element of string, and 0 is every element thereafter. Perhaps this was what you wanted -- it puts "0\0\0\0\0\0..." in the string after all, but I just thought I'd mention it. (If you do want to initialize every element in an array to something other than zero, either put every element in the initializer explicitly, or use something else like memset() or a for loop.)

Code:

//Allows user to input number
printf ("Please enter a number upto 255 digits long:\n");
scanf ("%s", number1);
printf ("Your first number is %\s.\n",number1);

printf ("Please enter another number upto 255 digits long:\n");
scanf ("%s", number2);
printf ("Your second number is %s.",number2);

scanf() is not the best way to read in a string. But I suppose you're not expecting anything other than digits, and you did say "up to 255 digits". Just as long as you know that that code is very fragile.

Code:

//Shifts the digits back so a bunch of 0's are not leading the number.
	for (b=255; result[b] != 0; b--){
	result[b] = result[b];
	}

The red code does nothing. You can leave the body of a for loop empty, so there's no need for it.

The for loop could also "underrun" the beginning of the array if there are no elements set to zero in result[0]..result[255]. Of course, that shouldn't happen, but it would if the string had overrun the end of the array.

Code:

memmove(result1,result+padding,255-b);

You do know what memmove() does, right? . . .

I guess that your for loops before the memmove()s are to determine the length of the strings. You can use strlen() in place of them. They're strings. If your compiler doesn't let you then you must have been passing the wrong argument to strlen(). This

Code:

	for (b=255; result[b] != 0; b--){
	result[b] = result[b];
	}
	padding = b+1;

	memmove(result1,result+padding,255-b);

should be able to be replaced with

Code:

	padding = strlen(result) + 1;

	memmove(result1,result+padding,255-b);

Code:

for(y=254; y>=0; y--) {

Why do you start at 254? Should that not be 255?

Code:

		sum = 0;
		sum += carry;

That's identical to

Code:

sum = carry;

Code:

sum += number1[y] - '0' + number2[y];

Why don't you subtract '0' from number2[y]?

Did you actually test this program? The not-subtracting-'0' bug would probably cause it to mess up . . .

Originally Posted by dwks

Code:

#include "stdafx.h"

But it may require turning off the "precompiled headers" setting, or changing some other settings regarding precompiled headers.

Indentation is horrible:

Code:

for (x= 0; number1[x] != 0; x++){
	number1[x] = number1[x];
	}
	padding1 = 255 - x;
	memmove(number1+padding1,number1,x);
	memset(number1, 0, padding1);

This looks like the memmove and memset are part of the loop. I prefer:

Code:

        for (x= 0; number1[x] != 0; x++){
	   number1[x] = number1[x];
	}
	padding1 = 255 - x;
	memmove(number1+padding1,number1,x);
	memset(number1, 0, padding1);

Of course, the for-loop in the above code can easily be replaced with:

Code:

   x = strlen(number1);

Also, there is no need to have x and z as variables for this and the next portion of code. Just use x for both loops. And just one "padding" variable.

I would however, as I suggested earlier, make sure that both numbers are the same length, so that you don't have to mess about if the strings aren't equal length. Do this by sticking '0' at the beginning of the string.

--
Mats

Originally Posted by matsp

But it may require turning off the "precompiled headers" setting, or changing some other settings regarding precompiled headers.

Hence "should".

Of course, the for-loop in the above code can easily be replaced with:

Code:

   x = strlen(number1);

I think I mentioned that, though perhaps it got lost in that mass of text . . .

I would however, as I suggested earlier, make sure that both numbers are the same length, so that you don't have to mess about if the strings aren't equal length. Do this by sticking '0' at the beginning of the string.

--
Mats

I believe that is what this code does:

Code:

//Shifts the numbers so they are aligned to the right
for (x= 0; number1[x] != 0; x++){
	number1[x] = number1[x];
	}
	padding1 = 255 - x;
	memmove(number1+padding1,number1,x);
	memset(number1, 0, padding1);

for (z= 0; number2[z] != 0; z++){
	number2[z] = number2[z];
	}
	padding2 = 255 - z;
	memmove(number2+padding2,number2,z);
	memset(number2, 0, padding2);

It makes number1 and number2 both contain the same number of digits, right-aligned. Without it, the math code would be somewhat more complicated.

Of course, it uses 0 where it should perhaps use '0', but the math code makes allowances for this.