#69710 - deltree - Wed Feb 01, 2006 2:13 pm
hi,
sorry about that beginner question again, it's about C and C++:
is it possible to write directly in binary on a variable ?
I know how to initialize a variable in hexa, but not binary.
how to do it ?
#69714 - tepples - Wed Feb 01, 2006 2:38 pm
C has no way to specify a literal constant in binary (base 2). It does have base 8 and base 16.
_________________
-- Where is he?
-- Who?
-- You know, the human.
-- I think he moved to Tilwick.
#69803 - Miked0801 - Thu Feb 02, 2006 12:31 am
foo = 01000101010b;
#69846 - keldon - Thu Feb 02, 2006 9:18 am
| Miked0801 wrote: |
| foo = 01000101010b; |
I did not know that. But since binary -> hex is so easy I (and most people who didn't know that) never cared anyway. But that will definately come in handy and make code much clearer now.
#69851 - Cearn - Thu Feb 02, 2006 10:17 am
| keldon wrote: |
| Miked0801 wrote: | | foo = 01000101010b; |
I did not know that. |
Neither did I. GCC (4.0.1) doesn't seem familiar with it either, all I get from that is
| Quote: |
| error: invalid suffix "b" on integer constant |
Can't find anything in its manual about it either. Are you sure it isn't just an extension of the compiler you're using?
#70660 - Miked0801 - Wed Feb 08, 2006 1:37 am
Truely? Hmmm - it was in my C book from way back (goes back and checks college text from way back.)
Well, well, well. It turns out I was remembering an old assembly book. In assembler, you can put h, o, b, d... on the end to change the base. My bad.
In C putting a 0 (zero) in front of a number tells the compiler it's octal, 0x for hex. Putting and l on the end of it is long, f - float, d = double, e = exponential (10e6 = 1 million). Nothing allows straight binary representation that I can see though.
#70661 - Miked0801 - Wed Feb 08, 2006 1:38 am
Which means you probably could "Inline asm" a db command with a b on the end for binary - but that's cheating...
#70667 - DekuTree64 - Wed Feb 08, 2006 2:34 am
Here's an ugly way to do it :)
| Code: |
#define BINARY_4BIT(b3,b2,b1,b0) ((b0)|((b1)<<1)|((b2)<<2)|((b3)<<3))
#define BINARY_BYTE(b7,b6,b5,b4,b3,b2,b1,b0) (BINARY_4BIT(b3,b2,b1,b0)|(BINARY_4BIT(b7,b6,b5,b4)<<4))
int one = BINARY_BYTE(0,0,0,0,0,0,0,1); |
_________________
___________
The best optimization is to do nothing at all.
Therefore a fully optimized program doesn't exist.
-Deku
#70720 - kusma - Wed Feb 08, 2006 11:24 am
there exists some nasty template-tricks to do it in c++ aswell... not very pretty, though.
#70721 - sgeos - Wed Feb 08, 2006 11:27 am
Packed bit fields?
-Brendan
#70758 - deltree - Wed Feb 08, 2006 4:39 pm
well... I think i'll have to work on translating binary to hex mentally;
#70760 - naleksiev - Wed Feb 08, 2006 4:49 pm
#70791 - Miked0801 - Wed Feb 08, 2006 8:03 pm
Trust me, being able to go back and forther from decimal to hex is a very necessary and useful skill to have. Once you do that, it's very simple to go from hex to binary and back. I'm to the point now where I can read many hex number without "translating" them in my head.
#70792 - kusma - Wed Feb 08, 2006 8:12 pm
| sgeos wrote: |
| Packed bit fields? |
the problem with that approach is that bitfields are allowed to be rearranged by the implementation.
#70796 - poslundc - Wed Feb 08, 2006 8:26 pm
| Miked0801 wrote: |
| Trust me, being able to go back and forther from decimal to hex is a very necessary and useful skill to have. Once you do that, it's very simple to go from hex to binary and back. I'm to the point now where I can read many hex number without "translating" them in my head. |
I agree with this statement. Hex is the new binary. Learn it, love it.
Dan.
#70911 - keldon - Thu Feb 09, 2006 3:14 pm
Try this:
| Code: |
#include<iostream>
using namespace std;
// Byte, shifted by n * 8
#define BN(b,n) (\
( \
( (b) & 1 ) | \
( (((b) >> 4) & 1 ) << 1 ) | \
( (((b) >> 8) & 1 ) << 2 ) | \
( (((b) >> 12) & 1 ) << 3 ) | \
( (((b) >> 16) & 1 ) << 4 ) | \
( (((b) >> 20) & 1 ) << 5 ) | \
( (((b) >> 24) & 1 ) << 6 ) | \
( (((b) >> 28) & 1 ) << 7 )\
) << (n) * 8 \
)
// Byte
#define B(b) (\
( (b) & 1 ) | \
( (((b) >> 4) & 1 ) << 1 ) | \
( (((b) >> 8) & 1 ) << 2 ) | \
( (((b) >> 12) & 1 ) << 3 ) | \
( (((b) >> 16) & 1 ) << 4 ) | \
( (((b) >> 20) & 1 ) << 5 ) | \
( (((b) >> 24) & 1 ) << 6 ) | \
( (((b) >> 28) & 1 ) << 7 )\
)
// Word
#define W(b1,b2) (\
BN((b1),1) | BN((b2),0 ) \
)
// Double Word
#define D(b1,b2,b3,b4) (\
BN((b1),3) | BN((b2),2 ) | BN((b3),1 ) | BN((b4),0 ) \
)
int main () {
cout << B ( 0x11111111); // prints 255
cout << BN ( 0x1,1); // prints 256
cout << D(0x0,0x0,0x1,0x0) << "\n"; // prints 256
return 0;
}
|
D is basically 4 bytes, making a double word. BN allows you to shift by a number of bytes - which you can tell by the output of BN (0x1, 1). I have chosen hexadecimal numbers simply so that you can write B(0x00001000) for simplicity and readability.
I tried doing this using long long's, but my gcc compiler does not like '(long long)0x076543210', so I am guessing Devkit Pro will not either.
EDIT: Added Word, and changed BD to D. Word is untested, but should work
p.s. I would have liked to use DB and DW, from the asm conventions. But I don't think there is one for double words
#71411 - keldon - Sun Feb 12, 2006 6:51 pm
Okay, 16 bit binary values work now, using long longs.
| Code: |
#define LW(b) (\
( (b) & 1 ) | \
( (((b) >> 4) & 1 ) << 1 ) | \
( (((b) >> 8) & 1 ) << 2 ) | \
( (((b) >> 12) & 1 ) << 3 ) | \
( (((b) >> 16) & 1 ) << 4 ) | \
( (((b) >> 20) & 1 ) << 5 ) | \
( (((b) >> 24) & 1 ) << 6 ) | \
( (((b) >> 28) & 1 ) << 7 ) | \
( (((b) >> 32) & 1 ) << 8 ) | \
( (((b) >> 36) & 1 ) << 9 ) | \
( (((b) >> 40) & 1 ) << 10 ) | \
( (((b) >> 44) & 1 ) << 11 ) | \
( (((b) >> 48) & 1 ) << 12 ) | \
( (((b) >> 52) & 1 ) << 13 ) | \
( (((b) >> 56) & 1 ) << 14 ) | \
( (((b) >> 60) & 1 ) << 15 )\
)
// ---
// HighByte LowByte
// --7654321076543210--
cout << LW ( 0x1111111111111111LL ) << "\n"; // prints 65535
cout << LW ( 0x0000000100000000LL ) << "\n"; // prints 256
|