Page 368 - Programming Microcontrollers in C

Numeric Encoding 353

{
if(*sp==’0') /* handle a zero input */
*sp=’0'+0xa;
*pq|=(*sp-’0')<<4 ;
}
else
pq|=0xf0; /* non digit, mark it */
if(isdigit(*(sp+1))) /* the next input */
{
if(*sp==’0') /* treat a zero */
*sp=’0'+0xa;
*pq|=(*(sp+1)-’0');
}
else
pq|=0xf; /* another non digit */
sp+=2; /* Increment the input data pointer, */
pq++; /* the output data pointer, */
i++; /* and the data count */
}
return i; /* length of the array */
}
Listing 7-1: Numeric Encoding Routine

All of the storage arrays in the EEPROM are of the type un­
signed int. Here the storage of both numeric data and alpha
data requires that every bit of every storage location be used so that
unsigned is the norm. In the coding routines, the data are passed in as
characters and the destination arrays are unsigned. Therefore, to aid
the local bookkeeping, the destination array pointer is immediately
assigned to a type char * and this pointer will be used to store the
encoded data.
The following program provides a simple test for the numeric
coding. This program has a serious problem though. All computers
configure memory in either a big endian or a little endian order. In
the big endian order, the most significant byte, 8 bits, of a 16-bit
address is given the smaller address and the least significant byte
goes to the larger address. The little endian order is just the reverse.