comparison is always false due to limited range of data type

[Dave]

hello...
I wrote a marco for saturation.
#define clip(x) (char)(x)<0?0(x)>255?255x));
and use this marco in the program like this...
char tmp=(char)clip((unsigned_int_16)(tmp1+tmp2));

which unsigned_int_16 will cast the (tmp1+tmp2) to 16 bit
unsigned int.

Each time when I compile this program, the gcc warning me the
"comparison is always false due to limited range of data type"

Where leads this problem?
Please tell me! Thanx~~~

Dave.

 

[Ben Pfaff]

I wrote a marco for saturation.
#define clip(x) (char)(x)<0?0(x)>255?255x));
[...]

Each time when I compile this program, the gcc warning me the
"comparison is always false due to limited range of data type"

Presumably `char' is unsigned, so the comparison (char)(x) < 0
is always false.

By the way, you seem to be over-fond of casts. You might want to
consider which casts are actually necessary and get rid of the
others.

 

[Randy Howard]

hello...
I wrote a marco for saturation.
#define clip(x) (char)(x)<0?0(x)>255?255x));
and use this marco in the program like this...
char tmp=(char)clip((unsigned_int_16)(tmp1+tmp2));

which unsigned_int_16 will cast the (tmp1+tmp2) to 16 bit
unsigned int.

Each time when I compile this program, the gcc warning me the
"comparison is always false due to limited range of data type"

What do you think x < 0 does on an unsigned data type?