fread returns strange values

I have this code segment meant to read the header from a standard ms paint bitmap.
It reads most of the stuff fine but returns strange values for 0xF9 0xC2 0xD8. (these are actually stored in a char data type but

Code:

#include <stdio.h>

int main(int argc, char *argv[])
{
  FILE *bmp;
  char *tmphead;
  int i;
  
  bmp = fopen("image.bmp","rb");
  if ( bmp == NULL )
    exit(0);

  fread(tmphead, 1, 0x35, bmp);
  
  for (i=0;i<0x35;i++)
    printf("%X %u\n",tmphead[i],tmphead[i]);

  fclose(bmp);
  scanf("%d");

  return 0;
}

and returns this

Code:

42 66
4D 77
E 14
37 55
2 2
0 0
0 0
0 0
0 0
0 0
36 54
0 0
0 0
0 0
28 40
0 0
0 0
0 0
FFFFFFF9 4294967289                   << ? this is not right (should be F9)
0 0
0 0
0 0
FFFFFFC2 4294967234                   << ? this is not right (should be C2)
0 0
0 0
0 0
1 1
0 0
18 24
0 0
0 0
0 0
0 0
0 0
FFFFFFD8 4294967256                   << ? this is not right (should be D8)
36 54
2 2
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0

The char is only supposed to hold 8 bits and this is not an 8 bit number.
I have compiled this with gcc distributed through Dev-Cpp package.

tmphead points nowhere (is uninitialized)

You probably want to make it an unsigned char as well - since that's why you get a load of f's at the beginning of the number - it's treated as a 32-bit signed number. Anything >= 0x80 in a signed char is negative, and when expanded to 32-bit will have 6 F at the beginning.

--
Mats

It declares a char as 32bit?
char is supposed to be 8 bit signed I thought.

He's trying to say that the format specifier you've chosen doesn't match what we think you're actually trying to read. Furthermore, you don't have any space allocated for your pointer, so you're reading into some random spot in memory. That's bad.


Quzah.

yeah I see

I always post before looking at my code

thanks

Also bear in mind that ALL integer types smaller or equal to int automatically get converted to int size. So if you pass a short or char to printf [or any other function that takes variable number of arguments] then the value will be int-sized. It doesn't matter if you use %c, %d, %u or some other format specicfier - the compiler may match up your arguments, but it doesn't look at the format string to determine what type to pass to printf.

Likewise, float becomes double, always.

--
Mats