The & Operator

[Albert]

Hello,

I'm having trouble understanding what K&R are saying: 'The bitwise AND
operator & is often used to mark off some set of bits, for example,

n = n % 0177;

sets to zero all but the low-order 7 bits of n.'

So in the end, what actually happens? What does the value of n become?

[Albert]

Re: The & Operator

And could someone give me an example of when the & operator is useful
or needed?

[Keith Thompson]

Re: The & Operator

"Albert" <albert.xtheunk nown0@gmail.com > writes:[color=blue]
> I'm having trouble understanding what K&R are saying: 'The bitwise AND
> operator & is often used to mark off some set of bits, for example,
>
> n = n % 0177;[/color]

You mean "n = n & 0177";
[color=blue]
> sets to zero all but the low-order 7 bits of n.'
>
> So in the end, what actually happens? What does the value of n become?[/color]

Think of the value of n as binary. 0177 is an octal constant; in
binary, it's 01111111. Ideally, n should be unsigned (bitwise operators
normally shouldn't be used with signed types).

For single-bit operands, the "&" operator is defined as:
0 & 0 == 0
0 & 1 == 0
1 & 0 == 0
1 & 1 == 1

Suppose the value of n (in binary) is 000101001110011 110001011010011 00
(32 bits). Then and-ing that value with 0177 yields:

000101001110011 110001011010011 00
& 000000000000000 000000000011111 11
--------------------------------
000000000000000 000000000010011 00

As you can see, this does just what K&R say it does: it zeros all but
the low-order 7 bits of n.

--
Keith Thompson (The_Other_Keit h) kst-u@mib.org <http://www.ghoti.net/~kst>
San Diego Supercomputer Center <*> <http://users.sdsc.edu/~kst>
We must do something. This is something. Therefore, we must do this.

[Markus Becker]

Re: The &amp; Operator

Albert <albert.xtheunk nown0@gmail.com > schrieb:
[color=blue]
> And could someone give me an example of when the & operator is useful
> or needed?[/color]

I guess you mean a real-world example ...

Think of some code-base that uses for example the following definition
for error values to set or return:

// Error-categories
#define CAT_FILEOP 0x0100
#define CAT_WHATEVER 0x0200

#define ERR_NOSUCHFILE (CAT_FILEOP|0x0 1) // '|' means OR where '&' means AND
#define ERR_FILEEXISTS (CAT_FILEOP|0x0 2)

// ...

#define ERR_OUTOFMEM (CAT_WHATEVER|0 x01)

Then you can use:

void check_error(int _err)
{
int cat,err;

cat = _err&0xf00; // mask out the specific error bits and get the category
err = _err&0x0ff; // mask out the category bits and get the error
hope("this helped"); // ;-)
}

Markus