Storing 2 numbers in a byte

I'd try something like this:




int lowBits = 9;
int hiBits = 12;

char Value = (lowBits & 0xf) | ((hiBits & 0xf) << 4);


With some casting maybe if the compiler is not happy about it

Check out the section on bit fields:
http://www.cs.cf.ac.uk/Dave/C/node13.html

Or, you can write a macro/function with the bitwise operators.

Personally, I wouldn't bother with either; unless it's for network compression or some other highly bandwidth limited purpose.

It's for saving bandwidth, thanks for the quick replies

If you are trying to save bandwidth and these values get passed a lot you might just go with a bit array instead of a byte. This allows you to send one number, two or more. Gives you he flexibility you might want in the future.

theTroll

You can try something like this:


union
{
unsigned char val;
struct
{
unsigned int hi: 4;
unsigned int lo: 4;
} parts;
};


Makes it a little easier and avoids bitshifting in favor slightly more readable code

If you want to save the values in a file or send them to another computer through a network or something, that isn't very safe though, since the compiler is free to choose how the bits get arranged. One compiler may put [font="Courier New"]hi[/font] in the upper bits and [font="Courier New"]lo[/font] in the lower bits, another compiler may do the exact opposite. Not a nice situation if you want portability across systems.

And then there's the deal of compilers just adding padding for the sake of performance (meaning not even alignment is guaranteed), although I don't think a compiler will add padding unless something would cross a byte/word/whatever boundary.