unsigned short
integer word or longword range: same as range of short or long
unsigned integer unsigned word or longword range: same as range of short or long
short signed word range: -32768...32767
unsigned short unsigned word range: 0...65535
OSErr signed word range: -32768...32767 (see Error Codes) LangCode signed word range: -32768...32767
Notes: The MC68xxx is a "Little Endian" meaning that the first byte in a 2-byte
word (the lower of the two bytes in memory) contains the most significant
part of the value (contrasted to "Big Endians," such as the 80x86 which
have the lowest byte of all data types as the least significant part). Thus, to
use a byte pointer to obtain a 16-bit word on the Mac, you could:
short theShort;
theShort = ( short)(bp[0] << 8) + bp[1];
The MC68000 (used in the original Mac and the Plus) is different from
the 80x86 with regards to odd-address word fetches (i.e., when you use a
MOV.W to read memory from an address which is not evenly-divisible by 2,
you get a CPU exception error and your program fails). Thus, you might
need to use code like the above on unformatted data area. Your compiler will
not watch for errors such as:
short * shortPtr, x;
bytePtr = mixedBuffer;
bytePtr++;
shortPtr = (short *)bytePtr; /* points to an odd address now */
x = * shortPtr; Ÿ/* System Error ID = 2 on old Macs */
Macintosh compilers will word align data structures so that you will not
get odd address errors by dereferencing a multiple word field that happens
to follow a 1 byte sized field. For example:
struct {
char x;
short y;
char z;
} foo;
sizeof() would return a value of 6 for this structure. The short field
starts on an even byte (word) boundary, and the structure itself is padded
to take up an even number of bytes.
Integers are either 2 or 4 bytes in size depending on your compiler and its
configuration.