C code prob continued - How a PC stores bytes?

[Jherek Carnelia]

Following on from an earlier post of mine (C code prob), I have to admit that I am baffled by the results returned by a small prog I have written.

I save a memory dump from Action Replay 3 (via WinUAE) to a file.

For instance, the values are - $aabb $ccdd $eeff $1122.

This is read into a PC (not an Amiga, where all is fine). Because of the differences between a PC and an Amiga, a check is done to determine whether or not the host is a PC or an Amiga.

If the host is a PC, the bytes are swapped, giving

$bbaa $ddcc $ffee $2211

Then I print the contents of the filebuffer as unsigned integers which gives

$bbaa $ddcc $ffee $2211 - so far, so good.

BUT, when I print the contents as bytes, I get -

$aa $bb $cc $dd $ee $ff $11 $22

Which is the original order before they were swapped.!

Got me stumped, I have to say!

Anyone know what's going on here? (the files are attached for those that care to try it)

[Zetr0]

looks to me the problem is in the print statements

Original code

	/* print the bytelist as short integers*/
	for(i = 0; i<BUFSIZE; i++) 
	{
		printf("$%04x ", *(buffer+i));  
        }
	printf("\n");
		
	/* print the bytelist as bytes */
	v_buffer = (unsigned char*)buffer;
	for(x = 0; x < 4; x++)
                printf("*(v_buffer+%d) = $%02x\n", x, *(v_buffer+x));

	getchar();

Try

	/* print the bytelist as short integers*/
	for(i = 0; i<BUFSIZE; i++) 
	{
		printf("$%04x ", buffer[ i ] );  
        }
	printf("\n");
		
	/* print the bytelist as bytes */

	for(x = 0; x < 4; x++)
                printf("bytelist byte = $%02x\n", buffer[ x ]);

	getchar();

I am unsure why you are using pointer references for print statements, its strange that I learned C when i was 16 coded it for the longest time, only to have used c++ for the last 6/7 years so its most likely that i forgot half of what i need to know LOL..

anyway i hope it helps.

[eLowar]

*(buffer+i) and buffer[i] are identical. Adding i (in pointer arithmetic) and dereferencing is exactly what operator[] does on pointers.

As for your output, keep in mind that little endian means "increasing numeric significance with increasing memory addresses" (thanks Wikipedia for a concise description). When you print whole words, you get the most numerically significant byte on the left (as humans write numbers), while when you print bytes by address you get it on the right (higher numeric significance means higher address).

It's late so I might not be thinking right, but I think that's it.

Edit: I just scrolled down a bit in the Wikipedia article that I stole the definition above from and it actually shows the same effect in the example for little endian storage. So there you go.

[Jherek Carnelia]

I'm glad I don't program in assembler on PCs. It must be a nightmare.

[eLowar]

I don't know. Higher address, higher value is pretty straightforward, isn't it? I mean it's not how you write numbers, but it means the exponent grows with the address offset.

[Npl]

Originally Posted by Jherek Carnelia

I'm glad I don't program in assembler on PCs. It must be a nightmare.

you typically dont have to care about endianess most of the times. Even if, its not really harder to deal with assembler than other languages.

x86 Assembler is a real nightmare tough, but for other reasons (lack of GPRs, nonorthogonal ISA ).