Binary constants

[valerij]

Hi,
what is the best way to hardcode into a C/C++ program binary
values/constants/etc. Since:

int a;
a=0b0111;

is wrong, how should I do it. I thought about converting them first to
hexadecimal or decimal, but what if you have a long string such as:

010100001010111 011010111010110 10 ... (64 digits long, for example for
__int64)

is there a way to handle this? And yes I have seen somewhere a very
smart macro that lets 0b01010 ... but it is limited to short strings.
Thanks,
Valerij

[Ondra Holub]

valerij napsal:

Hi,
what is the best way to hardcode into a C/C++ program binary
values/constants/etc. Since:

int a;
a=0b0111;

is wrong, how should I do it. I thought about converting them first to
hexadecimal or decimal, but what if you have a long string such as:

010100001010111 011010111010110 10 ... (64 digits long, for example for
__int64)

is there a way to handle this? And yes I have seen somewhere a very
smart macro that lets 0b01010 ... but it is limited to short strings.
Thanks,
Valerij

I (personally) would use hexadecimal mode. I do not think there may be
any "nice" solution, which doesn't call external function. 16 (+ 2)
characters for 64-bit number is not too much, I think it is no problem
(it is better than 64-bit binary representation) .

[Steve Pope]

valerij napsal:

Hi,
what is the best way to hardcode into a C/C++ program binary
values/constants/etc. Since:

int a;
a=0b0111;

is wrong, how should I do it.

That's because any constant starting with 0 is octal. (It
may interest people to know that in a simple expression like
i = 0, the "0" is octal and not decimal. Whereas in i = 1,
the "1" is decimal and not octal.)

You have your choice of octal, decimal, and hexadecimal.
Probably most systems programmers use hexadecimal. I agree
there's ocassions where binary would be useful, which is why
it is an option in some other languages.

Steve

[Salt_Peter]

valerij wrote:

Hi,
what is the best way to hardcode into a C/C++ program binary
values/constants/etc. Since:

int a;
a=0b0111;

is wrong, how should I do it. I thought about converting them first to
hexadecimal or decimal, but what if you have a long string such as:

010100001010111 011010111010110 10 ... (64 digits long, for example for
__int64)

is there a way to handle this? And yes I have seen somewhere a very
smart macro that lets 0b01010 ... but it is limited to short strings.
Thanks,
Valerij

Take a look at std::bitset:
http://www.sgi.com/tech/stl/bitset.html

#include <iostream>
#include <ostream>
#include <bitset>

int main() {
std::bitset< 64 bset(0xff);
std::cout << bset << std::endl;
}
/*
000000000000000 000000000000000 000000000000000 000000000001111 1111
*/

Note: a std::bitset< 64 is *not* the same type as std::bitset< 32 >
or any other size except 64.
Its not a true STL container but quite powerful in that it can easily
be converted to a string and has useful operators available. Its
input/output streameable too.

[Ivan Novick]

valerij wrote:

Hi,
what is the best way to hardcode into a C/C++ program binary
values/constants/etc. Since:

in C++ integer literals can only be in decimal, octal and hex.

010100001010111 011010111010110 10 ... (64 digits long, for example for
__int64)

In the C++ standard it says if an integer literal value can not be
stored in an int or long int than its undefined. So on most 32 bits
that would make a 64 bit integer literal illegal.

is there a way to handle this?

how about using string literals, and when your program starts up,
convert the binary string literals into __int64 at runtime?

-
Ivan
http://www.0x4849.net

[red floyd]

Steve Pope wrote:

valerij napsal:

>Hi,
what is the best way to hardcode into a C/C++ program binary
values/constants/etc. Since:

int a;
a=0b0111;

is wrong, how should I do it.

That's because any constant starting with 0 is octal.

0x....

[Steve Pope]

red floyd <no*****@here.d udewrote:

>That's because any constant starting with 0 is octal.

>0x....

No, 0x is hexadecimal, 0 without the x is octal.

Steve

[red floyd]

Steve Pope wrote:

red floyd <no*****@here.d udewrote:

>>That's because any constant starting with 0 is octal.

>0x....

No, 0x is hexadecimal, 0 without the x is octal.

Steve

I'm aware of that. However, to be precise, any octal constant starts
with '0'.

[Steve Pope]

red floyd <no*****@here.d udewrote:

>Steve Pope wrote:

>red floyd <no*****@here.d udewrote:

>>>That's because any constant starting with 0 is octal.

>>0x....

No, 0x is hexadecimal, 0 without the x is octal.

>Steve

>I'm aware of that. However, to be precise, any octal constant starts
with '0'.

Right, my mistake.

Any constant starting with 0 is either hexadecimal (if it starts
with 0x) or octal (otherwise).

Steve

[Harald van Dijk]

Steve Pope wrote:

red floyd <no*****@here.d udewrote:

Steve Pope wrote:

red floyd <no*****@here.d udewrote:

>>That's because any constant starting with 0 is octal.

>0x....

No, 0x is hexadecimal, 0 without the x is octal.

Steve

I'm aware of that. However, to be precise, any octal constant starts
with '0'.

Right, my mistake.

Any constant starting with 0 is either hexadecimal (if it starts
with 0x) or octal (otherwise).

0.1 is decimal. This is all not really relevant though: the original
code used a constant starting with 0b which is not allowed in standard
C++, so even if a compiler happens to accept it (as an extension), the
standard rules do not apply.