ANSI C Pre-processor - how to use constant suffix?

[xyz000]

Hi,

Suppose I have some code like this:

const unsigned int = 0x1234abcdU;

I want to achieve this with a macro. Something like:

const unsigned int = CREATE_UINT(0x1234, 0xabcd);

Optimistically I tried this:

#define CREATE_UINT(a,b) ((a << 16) | (b))U

The problem is that the "U" won't compile because of the brackets in the macro definition. I've tried many combinations of adding/removing brackets and of putting the U elsewhere but I still can't define a valid macro. Any ideas for a solution?

Thanks.

[GNU-Fan]

Everything that macros generate must be valid in the same sense as everything you would type directly must be valid.

Your makro writes this text:
((0x1234 << 16) | (0xabcd))U
which is not legal syntax, because the postfix "U" must follow a number (the figures written as text, not calculated).

My suggestion is to either write an ordinary function that returns, for example, an unsigned long.
Or you do a cast inside the macro, but you may get warnings for losing data.
(unsigned long*) a <<16 ...

[Rubberman]

As GNU-fan said, you cannot use the 'U' directive unless it immediately follows a literal value. You can do this, however:

#define CREATE_UINT(a,b) ((unsigned int)((a << 16) | (b)))